200 117 62MB
English Pages 288 [280] Year 2020
A FUNCTIONAL BIOLOGY OF STICKLEBACKS
FUNCTIONAL BIOLOGY SERIES General Editor: Peter Calow, Department of Zoology, University of Sheffield
A Functional Biology of Free-living Protozoa
Johanna Laybourn-Pariy
A Functional Biology of Sticklebacks R.J. WOOTTON Department of Zoology, The University College of Wales, Aberystwyth
U N I V E R S I T Y OF CALIFORNIA PRESS Berkeley and Los Angeles
©1984 R.J. Wootton University of California Press Berkeley and Los Angeles, California Library of Congress Cataloging in Publication Data Wootton, R.J. (Robin Jeremy) A functional biology of sticklebacks. Bibliography: p. Includes index. 1. Sticklebacks. I. Title. QL638.G27W663 1984 597'.53 ISBN 0-520-05381-8
Printed and bound in Great Britain
84-8864
CONTENTS
Series Foreword Preface Acknowledgements 1
Introduction
1
2
Spatial Distribution
4
3
Structure and Function
20
4
Feeding
32
5
Environmental Factors, Metabolism and Energetics
63
6
Growth and Production
83
7
Reproduction
103
8
Inter-specific Interactions
155
9
Population Dynamics
182
10
Ecological Genetics
193
11
Life-history Strategy in Sticklebacks
227
References
239
Index
261
For Siobhan and Sean, sceptical of sticklebacks
FUNCTIONAL BIOLOGY S E R I E S : F O R E W O R D General Editor: Peter Calow, Department of Zoology, University of Sheffield, England
The main aim of this series will be to illustrate and to explain the way organisms 'make a living' in nature. At the heart of this — their functional biology — is the way organisms acquire and then make use of resources in metabolism, movement, growth, reproduction, and so on. These processes will form the fundamental framework of all the books in the series. Each book will concentrate on a particular taxon (species, family, class or even phylum) and will bring together information on the form, physiology, ecology and evolutionary biology of the group. The aim will be not only to describe how organisms work, but also to consider why they have come to work in that way. By concentrating on taxa which are well known, it is hoped that the series will not only illustrate the success of selection, but also show the constraints imposed upon it by the physiological, morphological and developmental limitations of the groups. Another important feature of the series will be its organismic orientation. Each book will emphasise the importance of functional integration in the day-to-day lives and the evolution of organisms. This is crucial since, though it may be true that organisms can be considered as collections of gene-determined traits, they nevertheless interact with their environment as integrated wholes and it is in this context that individual traits have been subjected to natural selection and have evolved. The key features of the series are, therefore: (1) Its emphasis on whole organisms as integrated, resource-using systems. (2) Its interest in the way selection and constraints have moulded the evolution of adaptations in particular taxonomic groups. (3) Its bringing together of physiological, morphological, ecological and evolutionary information. P. Calow
PREFACE
Writing this book in the Functional Biology Series has allowed me to combine two of my major academic interests, research on the biology of the sticklebacks and teaching courses on theoretical ecology. The purposes of the book are twofold. The first is to demonstrate that the theoretical framework in ecology and evolutionary biology that has been developed, much of it over the past two decades, can be used to illuminate our understanding of the ways in which animals function in everyday life. The second is to show that the knowledge that can only be gained by a close and detailed study of a taxon will be required to test critically this theoretical framework. In writing this volume, I have been greatly helped by the advice of my colleagues. I would like to thank P. Calow, GJ. FitzGerald and F.A. Huntingford, who commented on the complete manuscript. M.A. Bell, J. Gee, H.Guderley,M.Milinski,D. Wharton and F.Whoriskey read sections of it. Their criticisms improved both the accuracy and clarity of the text. The mistakes of fact and interpretation that remain are my responsibility. I also thank Denise Long for her excellent drawings and R. Matthews for help with the word-processing. Finally, I thank my wife, Maureen, whose help and advice at all stages in the writing of this book were invaluable. R.J. Wootton Abefystwyth
ACKNOWLEDGEMENTS
All the Figures were freshly drawn, but I am grateful to the following for their permission to adapt copyright material. Figure 1.1 from Amer. Nat. 16, 775-84 (1976), Dr E.R. Pianka and the American Society of Zoologists. Figures 2.7 and 7.5 with permission from The Biology of the Sticklebacks', copyright: Academic Press, London. Figure 4.1 from Neth. J. Zool. 28. 485-523 (1978), Dr G.Ch. Anker and E.J. Brill, Leiden. Figure 4.2 from Behaviour IS, 284-318 (1960), Dr B. Tugendhat and E.J. Brill, Leiden. Figure 4.3 from Z. Tierpsychol. 45, 373-88 (1977), Dr M. Milinski and Paul Parey (Berlin). Figure 4.4 reprinted by permission from Nature 275, 642-44, Copyright © The Macmillan Press Ltd 1978 and Dr M. Milinski. Figure 4.5 from J. exp. Mar. Biol. Ecol. 25, 151-58 (1976), Dr R.N. Gibson and Elsevier Biomedical Press. Figures 6.2-6.4 from/. Fish Biol. 20,409-22 (1982), Fisheries Society of the British Isles. Figure 7.3 R.A. Moore. Figure 8.1 from Behaviour 10, 205-37 (1957), E J . Brill, Leiden. Figure 10.3 from Amer. Nat. 117, 113-32 (1981), Dr M.A.Bell and© 1981 The University of Chicago. Figure 10.5 from Evolution 33, 641-48 (1979), Dr D.W. Hagen and the Editor, Evolution. Figure 10.6 from J. Fish. Res. Bd Canada 26,3183-208, Fisheries and Oceans, Canada.
INTRODUCTION
The past two decades have seen a proliferation of theoretical studies in the related fields of ecology and evolutionary biology. Studies of optimal foraging, demography and life-history strategies spring quickly to mind. A strengthening of the theoretical framework of these fields is welcome, but it does bring with it an attendant danger. Mathematical equations can have a seductive elegance, simulation models developed on a computer a persuasive productivity. The danger develops when these theoretical developments are not tested by encounters with the often recalcitrant living organisms whose biology the theories should illuminate. The theoretical developments lose touch with the biological reality (Stearns, 1976,1977). Even when theories are tested, the diversity of living organisms is such that it may be possible to select a suitable example which seems to confirm the favoured theory. A better test of the adequacy of the theoretical framework is to apply it to organisms whose biology is already well known. How well can the framework be used to interpret and understand the life histories of active, living, evolving organisms coping with a difficult and changing world? Such an approach is also important in showing students that the theoretical framework can profitably be used to understand the biology of real organisms. Often students find it difficult to relate the theory to what they see living organisms doing. An animal can be viewed as a system that converts the energy and materials in its food into offspring. An input, food, is "mapped' into an output, the progeny (Figure 1.1) (Pianka, 1976). The success of the animal is measured by the number of its offspring that survive to reach their sexual maturity. This success has to be achieved in the face of possible shortages of food, living space and time in a potentially hostile physical and biotic environment. Although the description of the life histories of animals has always formed a significant part of zoology, the quantitative analysis of life histories is a recent development. This analysis has two components: theoretical studies of the consequences and the adaptive significance of patterns of survivorship 1
2
IN
Introduction
FORAGING TACTICS
OUT
Figure 1.1: Organism as an Input-Output System 'Mapping' F o o d into Progeny. Modified from Pianka (1976)
and reproduction and empirical studies which attempt to identify the causal factors responsible for the patterns observed in nature. Two sets of causal factors are sought, those responsible for the evolution of the observed life history, the ultimate factors, and those responsible for the day-to-day control of the components of the life history, the proximate factors. Unfortunately few species are studied in detail at each stage of their life history. Zoologists frequently study species which clearly illustrate some problem of immediate interest, while the other aspects of the biology of the species fade into the background. The significance of the features under study for the pattern of survivorship and reproduction of the species may even be ignored. A satisfactory analysis of the life history of a taxon also demands that the effects of experimental manipulations of environmental factors can be observed on the life history characteristics. Such manipulations are difficult or impossible for many species. A species that can be studied at each stage of its life history, both in the field and under experimental conditions, is the common, euryhaline, teleost fish, the threespine stickleback, Gasterosteus aculeatus. This species came into prominence because of a classic series of ethological studies on the reproductive behaviour of the male (Tinbergen, 1951), but other aspects of its biology command attention. It is geographically widespread and is variable physiologically, morphologically and in its life-history characteristics. This variability makes the stickleback an ideal subject for the analysis of life-history patterns. The other species in the stickleback family (Gasterosteidae) (Table 1.1) have been less studied, but they provide useful comparisons which help clarify the probable adaptive significance of some of the life-history characteristics of G. aculeatus. Although the biology of this family has been extensively reviewed
Introduction
3
Table 1.1: Species in the Stickleback Family (Gasterosteidae, Teleostei) Apeltes quadracus
The fourspine stickleback
Culaea inconstans
The brook stickleback
Gasterosteus aculeatus
The threespine stickleback
Gasterosteus wheatlandi
The blackspotted stickleback
Pungitius pungitius
The ninespine stickleback
Pungitius platygaster
The Ukrainian stickleback or small southern stickleback
Spinachia spinachia
The fifteenspine stickleback or the sea stickleback
(Wootton, 1976), no attempt was then made to interpret this biology in the context of a theoretical framework. The purpose of this book is to use the wealth of information that has been collected on the sticklebacks to illustrate how a picture of these fish, as they function in the real world, can be created by integrating both empirical data and the theoretical framework now available. Hopefully, such an approach will be applied to other well-studied groups of animals and plants, so that the symbiotic relationship between theoretical advances and empirical studies will be strengthened in ecological and evolutionary biology. Using the input-output model shown in Figure 1.1 as its major structural theme, this book describes how sticklebacks use their spatial habitat, food supply and time during their life and the patterns of growth, survivorship and reproduction that result. The successful 'mapping' of resources into progeny is mediated by the time and energy budgeting of the stickleback, budgeting which is controlled by physiological mechanisms within the constraints set by morphological design. The success of this mapping can be judged by the abundance and geographical distribution of the sticklebacks. Interactions of sticklebacks with other species, especially predators, parasites and competitors, influence the patterns of growth, survivorship and reproduction. But such interactions can also lead to evolutionary changes and the sticklebacks prove to be unusually suitable material for an analysis of the action of natural selection and other evolutionary processes on a vertebrate taxon. In a final chapter, the life history patterns of the sticklebacks are discussed in the context of theoretical studies on the evolution of life-history strategies.
SPATIAL DISTRIBUTION
Introduction Spatial distribution of the sticklebacks has several components: the global distribution of the family, the global distributions of the individual genera and species, and the local distributions of individual populations. A related factor is the extent to which the spatial distribution of populations reflects migrations between different habitats and localities. At one extreme, the global patterns reflect the evolutionary and adaptational history of the family and its species, whereas, at the other end of the scale, local distributions depend on the behavioural choices made by individual fish during their ontogeny. These behavioural choices are constrained by the physiological capacity of the fish to adapt to new environmental factors (Chapter S). Global Distribution of the Family The geographical range of the family is immense. Representatives are found throughout much of the northern hemisphere above about 35°N, with the exception of the continental heartland of Asia. In North America, Gasterosteids occur from the Great Lakes northwards, extending further south on both the eastern and western seaboards. Two species also occur in Greenland. The family is found in most of western Europe, and as far east as the Caspian and Aral Seas and as far north as the northern coastline of Siberia. In Asia, their distribution is restricted to coastal areas and islands southwards to Korea. Over large areas of this distribution two or more species are sympatric. This sympatry reaches a peak in north-eastern America, where, in the area of the St. Lawrence Estuary, the ranges of five species overlap. North America also holds the highest number of genera: Gasterosteus,Pungitius, Culaea and Apeltes. Three genera are found in Europe: Gasterosteus, Pungitius and Spinachia. Whereas in Asia only two genera occur, Gasterosteus and Pungitius. Because the evolutionary relationships 4
Spatial Distribution
5
between these genera are not understood, the events which have yielded this global pattern are not known. The teleost family to which the Gasterosteids are most closely related, the tube-snouts (Aulorhynchidae), is found only in the Pacific Ocean (Wootton, 1976). Global Distribution of the Genera Gasterosteus. The best known of the sticklebacks, G. aculeatus, is found on all three northern continents, and on all three its distribution is essentially coastal (Figure 2.1). In western North America, it is found along the Pacific coastline from Alaska to southern California. In the east it runs from Hudson's Bay and Baffin Island to Chesapeake Bay, and penetrates down the St. Lawrence Estuary to Lake Ontario. On the Pacific coast, the southern populations are confined to fresh water, but on the Atlantic coast, freshwater populations are rare south of Maine. It is not found along the northern coastline of North America, nor does it penetrate into the interior of the continent. In Asia, it is also absent from the northern coastline, but occurs in the Bering Straits and reaches as far south as Japan and Korea. Again there is no penetration into the continental heartland. Its distribution in Europe includes Iceland and as far north as the White Sea. In the south it occurs in Spain, Sardinia, most of Italy and parts of the Balkans peninsula. It is also found along the coast of the Black Sea and northwards into Poland and western Russia. It is not found east of the Black Sea nor in the Volga River. Its deepest penetration into Europe probably occurs down the Rhine, along the Dneister and Dneiper in south-east Europe, and along the Elbe, Oder and Neisse in north-eastern Europe. In Greenland, it is found on both the eastern and western coastlines up to about 70°N. Throughout this global distribution, there is a complex pattern of polymorphisms, which will be considered in Chapter 10. A second species of Gasterosteus, G. wheatlandi, is endemic to a relatively small area of the Atlantic seaboard of North America, including Newfoundland, the mouth of the St. Lawrence Estuary and as far south as New Jersey. Throughout this range it is sympatric with G. aculeatus (Figure 2.2). Pungitius. This genus is also found on all northern continents (Figure 2.3), and although sympatric with Gasterosteus over large areas, has a distinctive global distribution, whose 'centre of gravity' is more
Spatial Distribution
Spatial Distribution
9
northerly. In North America P. pungitius is found in coastal regions of Alaska and along the northern coastline of the continent, on Baffin Island, then around Hudson's Bay and the coast of Labrador, reaching as far south as New Jersey. It is also found deep in the continent, with a distribution stretching north-west from the Great Lakes. Throughout much of this continental distribution it is sympatric with Culaea. It is not found on the Pacific coast south of Alaska. In Asia its distribution is entirely coastal. Three distinct sub-species are recognised within this Asian distribution: P.p.pungitius occurs on the northern coastline of Siberia and the Kamchatka Peninsula and as far south as Japan; P.p.sinensis ranges from the southern coastline of Kamchatka to the Yangtse River and includes Japan in its range; P.p.tymensis is restricted to the islands of Sakhalin and Hokkaido and the west coast of the Sea of Japan. Over much of this range, Pungitius is sympatric with Gasterosteus. There is a continuous distribution of P. pungitius along the northern coastline of the Eurasian continental mass, which contrasts with the absence of G. aculeatus along this coastline. In Europe, P. pungitius ranges from the Arctic Ocean to the River Loire in France. The distribution is essentially coastal, so in Eurasia P. pungitius does not penetrate into the centre of the landmass as it does in North America. Throughout Europe there is sympatry with G. aculeatus. A second species of Pungitius, P. platygaster, is found in the river systems that drain into the Black, Caspian and Aral Seas. Culaea. This is one of the two genera restricted to North America (Figure 2.4). Culaea inconstans, in contrast to all the other sticklebacks, has a distribution that is largely continental, not coastal. It is found approximately between latitudes 40 and 60°N, east of the Rocky Mountains, from the Great Lakes north to Great Slave Lake. It reaches the Atlantic seaboard in Hudson's Bay and in the St. Lawrence Estuary. Throughout almost all of this range, C. inconstans is sympatric with P. pungitius, and in the east with G. aculeatus. Apeltes. As shown in Figure 2.5, A. quadracus is restricted to the Atlantic coast of North America, from the Gaspe Basin in Quebec in the north to Virginia in the south. Over much of this range A. quadracus is sympatric with G. aculeatus, G. wheatlandi and P. pungitius. Spirtachia. The largest and most elegant of the sticklebacks, S. spinachia is entirely restricted to the coastal waters of western Europe, from the Bay of Biscay in the south to northern Norway (Figure 2.6). It does not penetrate into fresh water anywhere within this range.
10
Spatial Distribution
12
Spatial Distribution
Figure 2.6: Approximate Distribution of Spinachia spinachia
The Habitats of the Sticklebacks Although the degree of geographical sympatry between the members of the stickleback family is extensive, the spatial overlap between local populations of the different species will depend on the habitat preference of the species within a geographical area. Habitat Preferences of Gasterosteus Gasterosteus aculeatus populations show one of two types of spatial behaviour. Some populations are resident in fresh water throughout the year, whereas other populations are anadromous, migrating to the sea in autumn and then migrating back into rivers, salt marshes or tidal pools in the spring to breed. River systems may contain both resident and anadromous populations. The leisurely mode of swimming of the stickleback excludes it from fast-flowing waters so the species is typically
Spatial Distribution
13
found in the still or slow-flowing water of lakes, ponds, lowland streams and sheltered coastal bays. Within this constraint, it lives in a wide variety of habitats ranging from large freshwater lakes to small ponds and streams. In some lakes, its spatial distribution is largely littoral but in others it is limnetic. Indeed, in some lakes, both littoral and limnetic populations may coexist (Chapter 10). The anadromous populations that migrate to the sea usually stay inshore, but sticklebacks have been found at distances up to 500km from the shore (Jones and John, 1978). Cod caught at depths of 200m and 160km from the coast of northern Europe had sticklebacks in their stomachs (Brown and Cheng, 1946). In the Barents Sea, sticklebacks can be abundant in the pelagic zone, usually associated with patches of floating algae (Parin, 1968). Within its limited distribution, G. wheatlandi is found predominantly in coastal waters, although it does occur at a few freshwater sites. During late spring and summer in the Rivière des Vases, a small river in the St. Lawrence Estuary, G. wheatlandi was only captured near the mouth of the river, where the salinity varied between 3 and 20ppt. None was found upstream in fresh water, nor was any collected in freshwater pools in the locality, although they were in brackishwater pools. At the same locality, G. aculeatus and P. pungitius were found in the river mouth, in the fresh water upstream and in freshwater pools (Worgan and FitzGerald, 1981a). Habitat of Pungitius Pungitius also has anadromous and resident populations. Its habitat preference is similar to that of Gasterosteus — the two can often be collected in the same net — but Pungitius probably prefers areas of denser vegetation than Gasterosteus (Wheeler, 1969). Collections of both species in eastern England suggested that P. pungitius was better adapted to shallow, weedy, eutrophic waters that became depleted of oxygen, whereas G. aculeatus preferred more open, well oxygenated waters (Lewis et al., 1972). In Rivière des Vases, the distribution of P. pungitius both in the river and in the salt-marsh pools was similar to that of G. aculeatus and included both the freshwater reaches of the river and freshwater pools. Adults of both species of Gasterosteus and P. pungitius had left the pools by early July, although the latter species persisted in the river until the end of August before dispersing into the St. Lawrence Estuary (Worgan and FitzGerald, 1981a). In Lake Huron, P. pungitius occurred close to the bottom at depths between 10 and 15m during the day, but showed a tendency to move deeper at night (Emery, 1973). Breeding P. pungitius in Lake Huron
14
Spatial Distribution
were observed among rocks around the shore (McKenzie and Keenleyside, 1970), but in Lake Superior, sexually mature fish were collected over highly organic muds at depths between 20 and 45 m (Griswold and Smith, 1973). Habitat of Culaea Sympatric with freshwater populations of P. pungitius over a wide geographical area, C. inconstans also has similar habitat preferences. It lives in heavily weeded, cool waters, including streams and ponds, as well as the swampy margins and bogs of lakes. It is found in prairie lakes which may become relatively deoxygenated in winter when covered with ice and snow, but has behavioural and physiological adaptations to survive these conditions (Chapter 5). Habitat of Apeltes Most populations of A. quadracus occur in vegetated areas with calm water. Populations commonly inhabit brackish estuaries and lagoons, whereas fully marine populations are rarer and tend to occur in sheltered inlets or mudflats where there is the eel-grass, Zostera. There are also resident freshwater populations throughout most of its geographical range, but it is probably not found in fresh waters of granitic areas where the water has a very low salt content (Blouw and Hagen, 1981). In Rivière des Vases, A. quadracus extended further up the river than G. wheatlandi but did not enter the freshwater stretches, nor was it found in any of the salt-marsh pools in which the other three stickleback species were breeding. It persisted in the river until the autumn (Worgan and FitzGerald, 1981a). In a small freshwater stream in Connecticut, A. quadracus was only collected in stretches containing the plant Elodea and where the current was slack. When fish from this stream were given a choice, they showed a strong preference for the half of a tank that contained Elodea rather than the half containing Potamogeton, another plant from the stream (Baker, 1971). Habitat of Spinachia Spinachia spinachia is a fish of shallow, marine waters. Beds of Zostera, weed-covered rocks, and stands of fucoid seaweeds are its preferred habitats. Although it lives in water with a salinity as low as 5ppt in the Baltic Sea, it is the only stickleback species that does not penetrate into fresh water. Although geographically it is sympatric with G. aculeatus and P. pungitius, there is probably little or no spatial overlap during the breeding season.
Spatial Distribution
15
Migration in the Sticklebacks Resident freshwater populations of Gasterosteus, Pungitius and Culaea may show a migration from the deeper water in which they overwintered into the shallower waters of small streams and backwaters where they will breed, but the most interesting migration is that shown by the anadromous stickleback populations. These migrate downstream to the sea soon after the completion of the breeding season in summer, and overwinter in the sea. In spring, the sexually maturing adults migrate back on to the breeding grounds, which may be in brackish or fresh water. This migration is a movement between waters of different salinities. In some species both resident and anadromous populations are found in some river systems, with the anadromous population typically moving into the lower reaches of the system during thé spring to breed. There may be an area of the system in which both populations are present, giving the potential for a gene-flow between the resident and anadromous populations (Chapter 10). The migration means that the population exploits one habitat during spring and summer while it is breeding and the young fish are growing rapidly, and another habitat during late autumn and winter. The possible advantages of this migratory behaviour will be considered later in the context of the whole life history of the species (Chapter 11), but the physiological mechanisms that permit this exploitation of two habitats differing in salinity must be considered. The migration has two components: a change in behaviour so that the fish make directed movements either up or down a salinity gradient (discussed in Chapter 5) and a process of physiological adaptation to the changing salinity as the migration proceeds. Osmoregulation in Sticklebacks For sticklebacks, as for other teleosts in fresh water, the osmotic pressure and the concentrations of inorganic ions of the blood and other body fluids are higher than in the external medium. Consequently, the fish tends to gain water and to lose inorganic ions, so that to maintain a constant osmotic pressure and concentration of inorganic ions of its body fluids it must excrete the excess water and take up inorganic ions. In sea water the situation is reversed. Compared with sea water, the body fluids have a lower osmotic pressure and concentration of inorganic ions. The fish tends to lose water to its external environment and to take up inorganic ions, so to maintain a constant internal environment the fish must acquire water and excrete the excess ions.
16 Spatial Distribution The main organs for osmo- and ionic-regulation are the kidneys, the gills and in some circumstances the intestine (Figure 2.7).
by gills
urine
Figure 2.7: Direction of Flow of Water and Salts when a Euryhaline Fish is in Fresh Water and Sea Water. After Wootton (1976)
The trunk portion of the kidneys contains clusters of renal corpuscles, each of which consist of a Bowman's capsule with its associated glomerulus. Fluids filter from the tangle of blood capillaries that form the glomerulus into the lumen of the capsule, from where they pass into the kidney tubule via a neck region. From the neck region the filtrate passes down the first proximal and second proximal segments into the collecting tubule. All the collecting tubules from a cluster of corpuscles fuse and open into the ureter which leads to the urinary bladder. The epithelial cells of the proximal segments and probably those of the collecting ducts and ureter modify the composition of the filtrate by reabsorbing or secreting inorganic ions (Wootton, 1976). In fresh water, the urine produced by the kidneys is more dilute than
Spatial Distribution
17
the blood serum, indicating that ions have been reabsorbed from the filtrate during its passage to the bladder. In sea water, the flow rate of urine is reduced to conserve water and the urine is at the same concentration as the blood serum (Wendelaar Bonga, 1973). The gills present a large surface area to the external environment which allows for the efficient exchange of oxygen and carbon dioxide, but they also present a large area through which water and inorganic ions can pass. In the epithelium of the gills are cells that are specialised for the transport of inorganic ions, the chloride cells. When the fish is in fresh water, the chloride cells probably take up inorganic ions, especially Na+and CI - , against the concentration gradient, but when the fish is in sea water, the cells excrete the excess inorganic ions again against the concentration gradient. When the fish is in sea water, it drinks to replace the water it is losing. The water is absorbed through the walls of the alimentary canal, but some of the ions in the sea water are lost in the faeces and others are excreted via the chloride cells of the gills. An unusual situation arises in the sexually mature male adult, because the kidney of the male becomes modified to secrete the mucus that is used to glue the male's nest (Chapter 7). This transformation of the kidney probably reduces its effectiveness as an organ of osmoregulation, and in mature male G. aculeatus there is an increase in the rate of production of fluid from the intestine. This fluid is less concentrated than the blood serum and so provides one pathway by which the mature male can get rid of excess water while it is breeding in fresh water (De Ruiter, 1980). Hormonal Control of Migration The downstream migration usually takes place when the day length is shortening and the water temperatures are declining, whereas the upstream migration takes place as day lengths increase and water temperatures rise. The changing day lengths and temperatures provide cues which trigger changes in the central nervous system and endocrine organs to induce the migratory behaviour and the process of physiological adaptation to a changing salinity (Baggerman, 1957). Some studies on G. aculeatus suggest that the thyroid gland which produces the hormone thyroxine is important in the control of migration. Anadromous G. aculeatus early in their upstream migration had highly active thyroids, and the lowest activity was shown by fish in the sea, although the activity of the gland declined during the breeding season and in the post-breeding phase while the fish were still in fresh water (Honma et al., 1977). The production of thyroxine is controlled
18
Spatial Distribution
by a hormone released by the pituitary, and the thyrotropic cells in the pituitary were maximally developed in fish that were collected in the sea in spring just prior to their migration into fresh water (Leatherland, 1970). Baggerman (1957) found that anadromous G. aculeatus treated with thyroxine developed a preference for fresh water over brackish water, and the hormone may also help to regulate the metabolic changes associated with increased locomotion required during upstream migration (Honma et al., 1977). When G. aculeatus in fresh water were fed with thyroid tissue, they showed disturbances in their ability to regulate the chloride content of the blood, and if treated in sea water many fish died (Koch and Heuts, 1942). As yet these studies do not provide a coherent picture of the specific role of thyroid hormones in the control of migration of the stickleback and, given the general effects that thyroxine has on metabolism as opposed to any specific effects it may have on migration, the teasing apart of general and specific effects will demand careful experimental studies. A pituitary hormone that seems crucial in the ability of anadromous G. aculeatus to move into fresh water in spring is prolactin. Injections of prolactin significantly reduced the mortality in fish that were transferred from sea water into fresh water in early winter. After the transfer, injected fish were able to produce a dilute urine and better regulation of the concentrations of Na+and CI" in the blood, abilities which uninjected fish also showed during the spring at the time of their normal migration into fresh water (Lam, 1972). The cells in the pituitary that produce prolactin seem to be most active in spring and least active in winter in anadromous fish, but in G. aculeatus permanently resident in fresh water, these cells are active throughout the year (Leatherland, 1970; Benjamin, 1974). Prolactin accelerated the changes in the structure of the epithelium cells of the renal tubules of migratory G. aculeatus that took place when the fish were transferred from sea water to fresh water (Wendelaar Bonga, 1976). Prolactin also reduced the net loss of Na+ and Cl"from the gill region, and induced an increase in the density of mucus cells on the gills (Lam, 1972). An increase in the thickness of the epidermis and an increase of mucus cells in the skin have been correlated with the activity of the prolactin cells of the pituitary (Wendelaar Bonga, 1978a). This stimulation of the activity of prolactin cells when the stickleback is in fresh water seems to be induced by the low Ca2+levels in the environment, rather than by low Na + or K + concentrations. The corpuscles of Stannius, small endocrine glands lying just behind and dorsal to the kidneys, probably play a role in the regulation of Ca2+ levels in the blood and so may also help
Spatial Distribution
19
to regulate the activity of the prolactin cells (Wendelaar Bonga, 1978a,b). Dispersal of the Sticklebacks With the ability of the stickleback to tolerate a wide range of salinities and the tendency towards migratory behaviour, the potential for dispersal along a coastline from river system to river system is high. The global distribution of both G. aculeatus and P. pungitius suggests that such dispersal has been extremely important, but the interpretation of the dispersal pattern in these two species is complicated by the marked polymorphism both in morphology and migratory behaviour that is shown by both species. This problem is taken up again in Chapter 10.
2
STRUCTURE A N D FUNCTION
Introduction Only a brief description of the morphology and anatomy of the sticklebacks will be given to provide the background for the other chapters. A more detailed description is given in Wootton (1976).
General Appearance Four of the five genera are similar in appearance. They are small spindleshaped fishes with a length-to-depth ratio that varies between about 4:1 and 5:1. They rarely exceed 100mm in length and are typically only half this. The exception is Spinachia, an elegant, elongated fish which can reach a length of 200 mm, with a length-to-depth ratio of about 11:1. All sticklebacks have a relatively slender tail stalk (caudal peduncle) with a truncate tail fin. The dorsal and anal fins are set well back along the body, but in front of the dorsal fin is a series of dorsal spines whose number varies from species to species. The paired fins, the pectorals and pelvics, have evolved distinct functions. Each pectoral fin has a narrow base set on the side of the body, but the fins are broad and rounded and are used in the normal mode of locomotion: the fish gently sculls itself along with synchronous beats of the pectorals. In contrast, the pelvic fins have lost their locomotory function, becoming defensive spines. The size and arrangement of the dorsal and pelvic spines vary between species although the basic arrangement is common. Another common feature of the family is the lack of the typical teleost scales. These have been replaced by bony plates or scutes, which are greatly modified scales forming a distinct row running down each flank of the body. Lateral plates are best developed in Spinachia, Gasterosteus and Pungitius, vestigial in Culaea and absent in Apeltes. Interand intra-specific variation in the number and pattern of the lateral plates provide one of the most intriguing functional and evolutionary problems posed by the sticklebacks (Chapter 10). 20
Structure and Function
21
Gasterosteus Gasterosteus aculeatus (Figure 3.1). This is the best known of the stickleback species and the onethat has attracted most attention from biologists, at least partly because of the ease with which it adapts to laboratory conditions and the hardiness with which it confronts those attentions. Its adult size is typically between 35 and 80 mm in total length, although in a few unusual populations smaller and greater lengths occur.
Figure 3.1: Gasterosteus aculeatus, a Completely Plated Morph from St. Lawrence Estuary, Canada. Total Length, 6 0 m m
Its common name, the threespine stickleback, is taken from the three dorsal spines that precede the dorsal fin. Each of these spines is separate from the others and the most posterior is separate from the fin, though this third spine is much shorter than the anterior two. The two pelvic spines are typically long and robust and are supported by a well developed bony pelvic skeleton, which forms a ventral shield on the belly. In some populations, however, the pelvic spines may be absent and the pelvic skeleton vestigial or absent (Chapter 10). A feature of G. aculeatus is the range of variation in the development of the lateral plates (Bell, 1976; Wootton, 1976; Hagen and Moodie, 1982). Three morphological forms or morphs are recognised on the basis of the number and arrangement of the lateral plates (Figure 3.2). The completely plated morph has a continuous series of plates which run from just posterior of the head to the tail stalk, where they form a distinct caudal keel on each side of it. The partially plated morph has an anterior series of plates, then a gap before a caudal series of plates, usually with a caudal keel. Only an anterior series of plates is present in the low plated morph, the rest of the flank being naked. In some populations no plates are present. Typically, the completely
22
Structure and Function
Figure 3.2: The C o m m o n Plate Morphs of G. aculeatus. T o p , Completely Plated; Middle, Partially Plated; Bottom, L o w Plated
plated morph has 30 to 35 plates in a lateral row, the low plated morph has from none to 14, and the partially plated morph has between 12 and 30. In much of the literature, the completely plated form is called trachurus, the low plated morph, leiurus and the partially plated morph, semiarmatus. These names do not have any formal taxonomic
Structure and Function
23
status but are convenient labels for the plate morphs. There is a correlation between the plate morphs and the tendency to migrate. Anadromous populations usually consist entirely or largely of the completely plated morph, although in Europe such populations often include the partially plated morph and small proportions of the low plated morph. The morph composition of freshwater resident populations is complex (Chapter 10), but populations consisting of only the low plated morph are common. The evolutionary status and significance of the three plate morphs has attracted much study (Chapter 10). Gasterosteus wheatlandi. This geographically restricted form of Gasterosteus (Figure 3.3) is morphologically similar to G. aculeatus, but there are strong basal cusps on the pelvic spines. The number of lateral plates is variable. Fish collected from Long Island usually have plates extending to the posterior part of the body whereas fish from the Gulf of Maine tend to lack posterior plates (Perlmutter, 1963). The most conspicuous differences between the two species are in aspects of their breeding biology (Chapter 7).
Figure 3.3: Gasterosteus
wheatlandi
f r o m St. L a w r e n c e E s t u a r y , C a n a d a . T o t a l
Length, 3 8 m m
Pungltius Although similar in general appearance to Gasterosteus, Pungitius (Figure 3.4) is easily distinguished by the row of relatively small dorsal spines whose number ranges from seven to twelve. These spines incline alternately to one side or the other of the midline, in contrast to Gasterosteus in which the spines stand vertically. The body is usually relatively slim with a slender tail stalk, so that Pungitius looks more delicate than Gasterosteus. The number of lateral plates can vary from none to 34. They are usually smaller than the plates in Gasterosteus, but they are best developed in P. pungitius sinensis.
24
Structure and Function
Figure 3.4: Pungitius pungitius from St. Lawrence Estuary, Canada. Total length, 57 mm
Culaea In general morphology, C. inconstans (Figure 3.5) is similar to Gasterosteus and Pungitius, although the body is deeper than in the slender Pungitius. The number of dorsal spines is usually between four and six, and these are comparable in size to those of Pungitius. Lateral plates are present but they are so tiny that only careful study has revealed them, so on routine examination C. inconstans looks unplated.
Figure 3.5: Culaea inconstans (Origin Not Known). Total Length, 50mm
Apeltes In many aspects of its breeding biology, A. quadracus (Figure 3.6) is an unusual stickleback (Chapter 7), but this does not extend to its general morphology. Its main features are a rather triangular body outline, an absence of any lateral plates, and between three and six relatively long dorsal spines. These dorsal spines incline to the right and left of the midline, and the last in the row is attached to the anterior of the dorsal fin.
Structure and Function
25
Figure 3.6: Apeltes quadracus from St. Lawrence Estuary, Canada. Total Length, 52 mm
Spinachia Spinachia spinachia (Figure 3.7) has a long thin body, with a long slender tail stalk. The head is also elongated, as a result of an elongation in the pre-orbital region. In the other sticklebacks, the jaw articulation lies approximately under the anterior border of the eye orbit, but in S. spinachia the articulation is well anterior to the eye so the elongation of the snout is not associated with an increase in the relative size of the mouth. There are about 15 spines in the dorsal row but they are small, and the pelvic spines are tiny and set well back along the body. There is a complete row of small lateral plates.
Figure 3.7: Spinachia spinachia from Cardigan Bay, Wales. Total Length, 9 5 m m
The Brain and Sensory System The brain of the stickleback reflects the relative importance of the sensory modalities in its life. Small olfactory bulbs lie on the anterior ventral surface of the telencephalon, the anterior portion of the forebrain. Most of the fibres from the olfactory bulbs project into the ventral part of the telencephalon, which consists of paired hemispheres. Behind the telencephalon, the diencephalon consists of the epithalamus, the thalamus and the ventral hypothalamus. A dorsal optic tectum
26
Structure and Function
forms a large part of the mid-brain, its size correlating with the importance of vision. The cerebellum, lying posterior to the optic tectum, probably controls co-ordination of movement and the maintenance of postural equilibrium. The posterior medulla oblongata merges into the spinal cord. Behavioural and anatomical evidence suggest that vision is the dominant sensory modality. The retinal surface accounts for about 3.5 per cent of the total surface area of the body and the retina contains both rods and cones. The sensitivity of the cones to colour can be correlated with the spectral absorption of natural water. Inshore and inland waters, the characteristic habitats of sticklebacks, are green, yellow-green or orange-brown, depending on the quantities of chlorophyll and products of the decay of vegetation. In G. aculeatus, three types of cone are present, one sensitive to blue, one to green and one to red light; the absorption maxima for the visual pigments are 452, 529 and 604nm, respectively. The cones are arranged in a well defined rectangular mosaic in the retina. In Spinachia, four types of cone are present, sensitive to blue, dark green, light green and orange (Loew and Lythgoe, 1978; Lythgoe, 1979). Behavioural studies on G. aculeatus have shown peaks of sensitivity to light of 510 and 594nm in females and 502 and 594 nm in males (Cronly-Dillon and Sharma, 1968). In addition to good colour vision, sticklebacks also have good form vision (Meesters, 1940). The epithelium of the two olfactory organs accounts for only 0.4 per cent of the surface area of the body. Although there is only a single opening to each olfactory organ, the functional morphology indicates that the flow of water over the sensory epithelium is unidirectional so that incoming and outgoing currents do not get mixed up (Theisen, 1982). Taste receptors are present in the mouth and pharynx. Olfaction seems to play little part in feeding, but olfaction and chemoreception may play a more important role in reproduction (Chapter 7). The role, if any, of sound reception in the life of sticklebacks is not known. Sticklebacks may produce sounds with their spines (Fish, 1954), but early attempts to condition sticklebacks to sound were unsuccessful (Westerfield, 1922).
Endocrine Organs Two types of endocrine organ can be recognised. Neuroendocrine
Structure and Function
27
organs represent parts of the nervous system that have become modified to secrete neurohormones; a variety of other organs form the other component of the endocrine system. At the core of the endocrine system is the pituitary organ, lying just under the brain, which acts as the primary link between the central nervous system and the endocrine system. Two components of the pituitary can be recognised. One is the neurohypophysis, which develops from an outgrowth of the floor of the brain. The function of the neurohypophysis in the stickleback is still unclear although it has been implicated in osmoregulation. A second neuroendocrine organ, the pineal complex, lies in the roof of the brain (van Veen et al., 1980). This complex also contains photoreceptor cells and may be important in the control of the timing of reproduction by the photoperiod (Chapter 7). A third neuroendocrine organ, the urophysis, lies at the posterior end of the spinal cord, but again its functional significance in the stickleback is not understood. The second part of the pituitary, the adenohypophysis, consists of several types of glandular cells whose secretions are released into the blood system and control either directly or indirectly most anabolic and catabolic processes including growth and reproduction. In the anterior part of the pituitary are the prolactin cells whose secretory product, prolactin, is important in osmoregulation for sticklebacks in fresh water (Chapter 2) and may also be important in the regulation of male parental behaviour (Chapter 7). Dorsal to the prolactin cells lie the cells that produce adrenocorticotrophic hormone (ACTH), which controls the activity of the inter-renal endocrine tissue. In the middle region of the pituitary, three types of glandular cell have been identified. One type secretes the thyroid stimulating hormone (TSH), a second secretes somatotropic (growth) hormone (SH), and a third secretes gonadotrophs hormone (GTH), although some evidence suggests that two types of cells perhaps secreting two different GTHs may be present (Slijkhuis, 1978). The secretory activity of these cells in the pituitary is probably controlled by neurohormones produced in the brain, which either inhabit or stimulate the secretion of specific pituitary hormones. Although the general pattern of the control of the pituitary hormones in sticklebacks is similar to that of other teleosts, few of the details are known. The major endocrine target organs for the pituitary hormones are the inter-renals, the thyroid and the gonads. Inter-renal tissue lies in the paired head kidneys and it is assumed to be homologous with the adrenal cortex tissue of terrestrial vertebrates. Thyroid tissue lies
28
Structure and Function
diffusely on the dorsal surface of the ventral aorta in the pharyngeal region ventral to the oesophagus. This endocrine organ has been implicated in osmoregulation and migratory behaviour in sticklebacks (Chapter 2), but probably also regulates metabolic activity. The gonads — paired testes or ovaries — lie in the abdominal cavity, dorsal to the alimentary canal. In males, the interstitial tissue is the endocrine component of the testes, producing androgens. In females, it is probable that the follicle cells which surround a maturing oocyte produce the female sex hormones, oestrogens and progesterones. Two endocrine organs which are probably concerned with the regulation of the concentrations of ions in the body fluids are the corpuscles of Stannius and the ultimobranchial organ. The former lie on the posterior dorsal surface of the kidneys, and the latter is found in the septum between the abdominal cavity and the sinus venosus just ventral to the oesophagus. Studies on the function of these organs in the stickleback are only just beginning (Chapter 2). The endocrine component of the pancreatic tissue occurs in tissue that lies near the pyloric sphincter, which separates the stomach from the intestine. By analogy with other vertebrates, the pancreas probably regulates carbohydrate metabolism, growth and protein synthesis, although analysis of the carbohydrate content of the tissues of the stickleback suggests that carbohydrate metabolism probably forms a relatively small component of the metabolic activity. Other endocrine tissue is found in the kidneys and probably in the alimentary canal. Although the small size of the stickleback has proved an advantage in many experimental studies, it is a disadvantage in studies on endocrinology, and most of the functional significance of the endocrine organs discussed has been inferred from studies on the histological and histochemical properties of the organs.
Muscle and Movement The trunk musculature of the sticklebacks lacks red muscle, which in teleosts is used for long periods of sustained swimming. Such red muscle operates aerobically. The trunk musculature consists entirely of white muscle which operates anaerobically (Boddeke et al., 1959; Kronnie et al., 1983). This white muscle is used for short rapid bursts of swimming, but an oxygen debt is built up so that the muscle fatigues rapidly in comparison with red muscle. Because they lack red muscle, sticklebacks have been classified as 'sprinters' in terms of their loco-
Structure and Function
29
motory capabilities (Boddeke et al., 1959), although this description is perhaps flattering. Sticklebacks are not highly active swimmers. Their normal form of locomotion is a leisurely sculling with the pectoral fins — the labriform mode of locomotion (Lindsey, 1978). Associated with this mode of locomotion is a well developed pectoral skeleton. At times of excitement, when escaping from a predator, chasing a rival or approaching a female, the stickleback switches to a more typical teleost mode, carangiform locomotion. The posterior region of the body is thrown into waves which pass posteriorly to the tail. In sticklebacks, the relatively well developed post-cranial skeleton and the lateral plates tend to restrict the extent of the body that generates the locomotory wave. This carangiform swimming cannot be maintained for any extended period of time ; it is simply used for short rapid bursts of swimming. Sticklebacks spend much of their time hovering motionless in the water column. They have a well developed swim bladder which lies in the dorsal roof of the abdominal cavity. Shortly after hatching, a young stickleback fills this bladder by swimming to the surface and gulping air, but soon afterwards the connection between the bladder and the alimentary canal is lost. Thereafter, pressure changes in the gases in the bladder are controlled by the absorption and secretion of gases through its wall. The anterior portion of the bladder secretes gases into the bladder whereas the posterior portion, which is separated from the anterior portion by a diaphragm, can reabsorb them. Adjustments in the quantity of gas in the bladder allow the stickleback to maintain neutral buoyancy as the fish moves up and down in the water column (Fange, 1953).
Urinogenital System The kidneys and gonads lie dorsally in the abdominal cavity above the alimentary canal. The paired kidneys lie in the roof of the cavity and each kidney has two distinct components, an anterior head kidney linked by a strand of tissue to the posterior or trunk portion. This is also a functional division because the head kidney contains lymphatic, haemopoietic and endocrine tissue whereas the trunk portion has, as its primary function, the production of urine (Chapter 2). This posterior portion also has an important function in reproductively active males, for the glue used in the construction of the male's nest is synthesised by the cells of the kidney tubules (Chapter 7). The gonads
30
Structure and Function
consist of a pair of testes or ovaries and lie just below the swim bladder. Except in young immature fish, the gonads can be distinguished macroscopically, with the ovaries soon exceeding the size of testes; the visceral peritoneum that covers the testes contains many melanophores, with their black pigment. In sexually mature females, the ovaries expand almost to fill the abdominal cavity and this causes a visible distension of the abdomen.
Digestive System The alimentary canal forms a simple tube, with the oesophagus leading into a true stomach which is separated from the anterior intestine by a pyloric sphincter. The intestine leads to the rectum from which it is indistinguishable histologically (Hale, 1965). There are two lobes to the liver, with the larger right lobe reaching past the pyloric region of the stomach. The liver shows changes in size that are correlated both with the nutritional state of the fish and with the reproductive cycle (Chapters 6 and 7). As in other teleosts, it is probably the site at which the bulk of the material destined to form the egg yolk is synthesised (Chapter 7). Other organs associated with the alimentary canal are the pancreatic tissue and spleen. Well fed fish accumulate fat bodies around the alimentary canal, but these tend to become reduced or disappear altogether in poor feeding conditions or during winter.
Systematic Position of the Sticklebacks Similarities in anatomy, morphology, behaviour and ecology indicate that the sticklebacks form a natural, monophyletic family, but the evolutionary relationships between the members of the family are not known. Culaea, Pungitius and Gasterosteus may be closely related, with Apeltes and Spinachia not closely related to that group, nor to each other. Unfortunately, the relationship of the sticklebacks to other teleosts is even more obscure. They are obviously closely related to the tube-snouts, the Aulorhynchidae, but this family contains only two species and provides no real clues to the relationships between the tube-snouts and sticklebacks and other teleosts. Conventionally, the sticklebacks are classified in the order Gasterosteiformes whose other members include fishes such as the pipefishes, seahorses and other bizarre marine forms (Wootton, 1976). In any studies of the
Structure and Function
31
sticklebacks it should be remembered that they represent an evolutionary sideline of the teleosts and so generalisations from sticklebacks to other teleosts should be made carefully and critically.
FEEDING
Introduction Feeding behaviour determines the resources that the stickleback has available for investment in survivorship, growth and reproduction. It is the critical activity of the daily life of the fish. The details of the feeding behaviour determine both the qualitative composition of the diet and the quantity of food eaten. A feeding or foraging sequence can be broken down into a number of elements. The fish initiates a search, which eventually terminates with the discovery of a potential prey. This prey is approached and if it tries to evade capture it is pursued. If the prey does not escape, capture then ingestion follow. At each stage in this sequence after the detection of the prey, the fish may reject that prey and either turn its attention to another prey item or resume searching. The number of feeding sequences that are completed during a day will determine the quantity of food consumed, and the pattern of rejections and acceptances will determine the composition of the diet. A detailed analysis of the feeding sequence under a range of environmental conditions (including factors such as the size distribution and density of the prey, the density of the sticklebacks, and the temperature and light regimens), which also takes account of the physiological state of the stickleback, will allow the development of a predictive model of its feeding tactics.
Detection of Prey Sticklebacks are visual predators (Beukema, 1968; Wootton, 1976), so it is prey characteristics that provide visual cues that are important for the detection of prey. Important cues are size, movement, shape, colour contrast with the background and possibly oddity in movement, colour or shape. The distance at which a fish can first detect a potential prey is 32
Feeding
33
called the reaction distance. This distance will depend on the size of the image that the prey item projects on to the fish's retina, so that larger prey can be detected at a greater distance (Eggers, 1977). In clear water, G. aculeatus could detect lOmm-long Gammarus or Asellus at a distance of 440 mm although in turbid water the reaction distance was reduced to 260mm (Moore and Moore, 1976). A single Tubifex worm was detected at 150mm in every trial, and at 250300mm in 50 per cent of the trials (Beukema, 1968). In the tests with Tubifex, fish with the shortest reaction distances were those that had some eye deformity such as swelling or bulging. The reaction distance to Daphnia magna that measured 2.8mm from the head to the base of the caudal spine was 120mm for red and 102mm for pale-yellow daphnids (Ohguchi, 1981). Even casual observations suggest the importance of movement for the detection of prey. It is much more difficult to get a stickleback to feed on motionless food such as commercial fish food than on a bunch of wriggling Tubifex or enchytraeid worms. Spinachia spinachia preferred moving mysids to stationary ones unless the prey were moving very slowly (0.2 cms"1). Movement of the prey increased the frequency at which the S. spinachia attempted or completed feeding responses (Kislalioglu and Gibson, 1976b). In an early series of experiments on the visual responses of G. aculeatus, Meesters (1940) found that the fish responded more to a wavy thread than to a motionless straight thread, but when the straight thread was moved, the response increased to a maximum, then declined when the thread was moved at high speeds. The speed at which the response was maximal was 3 cms -1 . A similar optimum speed for the movement of prey was found for S. spinachia (Kislalioglu and Gibson, 1976b). When tested with swarms of ten Daphnia, G. aculeatus directed their first attacks towards the Daphnia that did not match the colour of their background. The Daphnia were either pale-yellow or red and were tested against a background that was either pale-yellow or red (Ohguchi, 1918). This importance of the contrast between the prey and its background was demonstrated when G. aculeatus were allowed to hunt for single Drosophila larvae that were white and almost motionless. If patches of white cloth were placed on the normally dark bottom of the experimental tank, the chance that the larvae would be detected by the fish dropped sharply and the reaction distance was much reduced (Beukema, 1968). Gasterosteus aculeatus feeding on corixids in a pond took a greater proportion of the corixids whose colour made them conspicuous against the substrate (Popham, 1966). Spinachia spinachia
34
Feeding
was found to prefer dark to light mysids, although the relative importance of the stimuli presented by the mysid was such that movement was as or more important than size and both were more important than colour, and shape was least important (Kislalioglu and Gibson, 1976b). Some evidence suggests that predators detect or select prey that differ in some respect from the majority of their kind. But in an experiment in which abnormally moving Daphnia were produced by amputation of parts of one or both antennae, no evidence was found to suggest that G. aculeatus consistently preferred the oddly moving prey (Milinski and Lowenstein, 1980). In contrast, G. aculeatus did turn their attention to Daphnia that were different in colour from other members of the swarm, even when the oddly coloured Daphnia were not conspicuous against the background (Ohguchi, 1978, 1981).
Foraging Behaviour in Experimental Studies on G. aculeatus Behavioural Organisation of Foraging If the prey is discovered on the substrate, the stickleback approaches and tilts its body, adopting a snout-down position while fixing the prey with its eyes. It seizes the prey with a rapid sideways twist and a snapping movement (Tugendhat, 1960a). High-speed photography has shown that the mouth opens with an explosive expanding movement, sucking the prey in to be grasped by the short but sharp teeth of the jaws (Figure 4.1). The snapping action is characterised by a very rapid opening of the mouth, immediately followed by a slower closure. If the prey is small, it is swallowed immediately, whereas if the prey is large it may be repeatedly grasped then spat out. This spitting action reverses the snapping because the mouth is opened relatively slowly but then closed rapidly. Since the prey are not chewed, this grasping and spitting may help to break up larger prey so that they can be easily swallowed. The speed at which the mouth opens has been found to vary with the size of the prey. Higher speeds are used when the prey is a relatively thick earthworm than when it is a slender Tubifex. Speeds of up to 283mms -1 have been observed (Anker, 1978). The probability that a foraging sequence ends successfully with the consumption of a prey depends on a number of factors including the hunger of the fish, the palatability and density of the prey, and the extent to which the foraging is disturbed.
Feeding
35
Figure 4.1: Gasterosteus aculeatus Feeding on a Worm. After Anker (1978)
Hunger. Typically, the hunger of sticklebacks is manipulated experimentally by depriving the fish of food for known periods of time. Two studies have analysed in detail the effect of hunger on foraging behaviour. In a series of experiments by Tugendhat (1960a, b), the sticklebacks fed on a dense population of Tubifex in a relatively simple feeding compartment, whereas Beukema (1968) studied sticklebacks feeding on single Tubifex in a relatively complex feeding compartment made up of 18 hexagonal, interconnecting cells. Despite the differences in experimental design, the basic results of these two studies were similar. Tudendhat (1960a) found that as the period of food deprivation was increased from 1 day up to 3 days, the number of completed feeding responses during a test lasting 1 h also increased. The total number of responses that were initiated by the fish visually fixing the prey was not affected by the length of the deprivation period, but the
36
Feeding
Figure 4.2: Time Course of Frequency of Completed Feeding Responses Made by Feeding on Tubifex after Three Periods of Food Deprivation. After Tugendhat (1960a)
G. aculeatus
average length of time spent on a response, whether completed or just initiated, was longer the shorter the deprivation time. Although the hungrier fish completed more feeding responses, they did not spend a greater proportion of the test-period feeding. Fish spent just over 20 min making feeding responses. Within the hour-long feeding session, these effects of deprivation were reversed. As the hour progressed, the number of completed responses declined although the number of initiated responses did not. The time course of the changes in the number of completed responses during the hour was complex (Figure 4.2). There was an initial rapid decline which was fastest for the fish that had suffered the longest deprivation periods, but then there was
Feeding
37
a slight recovery in the frequency of completed responses followed by a further slow decline. Thus hunger did not influence the total time spent feeding, but increased the frequency of completed feeding responses, especially during the first part of a feeding session. In fish that had been deprived of food for periods up to 88 h and then provided with ad lib Tubifex for 8 h, the total daily consumption was not related to the period of deprivation, but the proportion of the daily total that was consumed in the first hour increased with the period of deprivation (Beukema, 1968). Again, hunger affected the initial frequency of feeding rather than the final amount consumed. When a stickleback was allowed to search for a single Tubifex in a complex feeding compartment, a number of quantitative components of the foraging sequence were found to increase as the period of food deprivation was increased (Beukema, 1968). Increases occurred in: the proportion of grasped prey that were eaten, the proportion of discovered prey that were grasped, the number of feeding responses directed towards inedible objects, the number of bursts of search swimming, the distance swum and the number of cells in the feeding arena visited. This positive correlation between these disparate components of foraging is good evidence that hunger can be regarded as a unitary motivational tendency which can be manipulated by food deprivation and measured by one of these covarying components. A component of foraging behaviour that was not affected by the period of deprivation was the reaction distance. As the single Tubifex was discovered and eaten it was replaced, so that during the course of the 40 min test period the fish became progressively satiated and the components which were positively correlated with hunger decreased in value. The total time spent handling prey was not affected by the period of deprivation because, at the shorter deprivation periods, the frequency of handling the prey decreased but the mean length of time for which each prey was handled increased. Fish size was positively correlated with the frequency of cell visits, the proportion of prey that were grasped after discovery and the proportion of prey that were eaten after being grasped. These three components were all also dependent on the period of deprivation, which suggests that their relationship to fish size was mediated through hunger. Larger fish satiated more slowly than small fish (Beukema, 1968). If three Tubifex were present rather than one, there was no effect on the number of worms eaten in the 40 min period, although there was an increase in the rate of eating in the first few minutes of the period. The main effect of the higher prey density was to reduce the
38 Feeding proportion of prey grasped and the proportion eaten, so that each individual prey item had its risk of being eaten reduced by one-third. As the fish became more satiated, the proportion of prey grasped declined more than the proportion eaten, which suggests that the rejection of a particular prey tended to be made after it was discovered but before it was grasped (Beukema, 1968). These results were comparable to those obtained by Tugendhat (1960a). In both series of experiments, the hunger of the experimental fish was determined by the period of food deprivation prior to the feeding session and by the number of prey consumed during the session. Palatability. Acceptance or rejection of a prey item has a significant effect on subsequent foraging behaviour (Thomas, 1974, 1976, 1977). Sticklebacks were allowed to hunt for Tubifex worms in a long, linear feeding compartment. If the fish discovered a worm and ate it, then there was a reduction in the tendency of the fish to move from the area of discovery, and an immediate increase in the intensity of searching and in the frequency of snout-down approaches to the substrate. These changes in foraging behaviour led to 'area-restricted' searching. In contrast, if the Tubifex were rejected, there was an increased tendency for the fish to move directly away from the site of the prey and an initial decrease in the intensity of searching. This behaviour led to 'area-avoided' searching. Neither the outcome of an encounter with the prey, acceptance or rejection, nor the behaviour shown immediately after the encounter with the prey was totally dependent on the current encounter; they also depended partly on the previous feeding history of the fish. Nor could the changes in behaviour be accounted for solely in terms of an increase in the satiation of the fish during the test session. Thomas (1976) suggested that short-term positive and negative motivational after-effects were present after the acceptance or rejection of the prey, so that an acceptance positively influenced the probability that a subsequent prey would be eaten whereas a rejection negatively influenced that probability. These short-term effects of an encounter with a prey were estimated to have a time span of the order of 1-12 min. A relatively complex pattern of changes in the frequency of completed responses to prey was also seen in Tugendhat's experiment. In Thomas's experiments, the prey was always Tubifex, so the cues that caused the fish to accept or reject a particular prey are not clear. Beukema (1968) analysed the effect of two prey types on the behaviour of sticklebacks foraging in his system of hexagonal cells. The new prey types used were either Drosophila larvae or pieces of enchytraeid
Feeding
39
worms. The densities of prey were maintained at one Tubifex and two of the new prey, so that two prey types were always present. The larvae were smaller than the Tubifex and virtually motionless. They were less acceptable to the fish than the Tubifex, but the rate of eating the larvae increased significantly after their introduction up to a level that was either stable or declined slowly. The presence of the Drosophila had no effect on the rate of feeding on Tubifex. In fish that found the larvae acceptable, the only component in the foraging behaviour sequence that increased consistently was the proportion of prey that was discovered. The reaction distance to the larvae did increase as the number of larvae taken increased, and there was also an increase in the number of responses to white inedible objects. Beukema suggested that the increase in the proportion of larvae discovered had two components: the fish learned fairly rapidly to respond to white objects in general and then learned, at a slower rate, to distinguish between inedible white objects and the larvae. The enchytraeid pieces were highly palatable to the fish, and nearly every piece that was grasped was eaten. As the number of enchytraeids eaten increased, the risk to the Tubifex declined. The increase in the number eaten was caused by an increase in the proportion that were discovered and an increase in the rate at which cells in the feeding compartment were visited. There was also an increase in the rate of response to inedible white objects, but a reduction in the rate of response to non-white inedible objects. For both types of new prey, it took many feeding experiences before the fish reached their maximum ability to discover the new prey. In this two-prey experiment, the proportion of Tubifex and Drosophila that was eaten after being grasped declined as the fish's stomach filled, though the proportion of Tubifex eaten was consistently higher than the proportion of Drosophila. The proportion of enchytraeids eaten was high at all times. During a session, the Drosophila tended to be rejected earlier than the Tubifex, which was rejected earlier than the enchytraeid. The effect of the presence of Drosophila on the risk of the Tubifex being eaten was merely the equivalent of an increase in the density of the Tubifex. In contrast, the presence of the enchytraeid caused a significant reduction in the risk to the Tubifex. This risk fell to low levels within the first 5 min of a feeding session as the proportion of Tubifex discovered, grasped and eaten fell. Even in subsequent sessions in which no enchytraeids were present, the risk to the Tubifex did not completely recover and the fish's responsiveness to white objects remained high. The fish had learned to associate white objects with highly palatable prey (Beukema, 1968). Earlier, Meesters
40
Feeding
(1940) had shown that sticklebacks showed a preference for snapping at drawings of solid objects after they had been eating a solid prey, whereas, after eating a thin prey, they tended to snap at drawings of lines and crosses. Thus learning the characteristics of the prey plays a significant role in the foraging behaviour of the stickleback. The relative values of the three prey items to the sticklebacks are not known, so it is not clear whether the strong preference for enchytraieds was because they were a food of high quality. Density of Prey. Beukema's experiments used very low densities of prey, so that at a maximum there were only three prey in an area of 187 dm 2 . In Tugendhat's experiments the densities of Tubifex were much higher, but these benthic worms showed little or no evasive behaviour when approached by the stickleback. Observations on sticklebacks preying on swarms of free-swimming Daphnia suggest that high, local concentrations of the prey can influence the behaviour of the fish. One of the functions suggested for schooling behaviour which results in relatively high local densities of the schooling animal is that it reduces the effectiveness of a predator attacking the school and so improves the chances of each prey surviving the predatory attack (Bertram, 1978). A single stickleback attacking a swarm of Daphnia was observed to prey preferentially on stray Daphnia. This preference for strays tended to increase as the stickleback became satiated. When confronted with a dense swarm, the fish usually stopped in front of it and then often withdrew without attacking (Milinski, 1977a). This effect of swarms of Daphnia on the behaviour of the stickleback was investigated by presenting the fish with different densities of Daphnia confined in small tubes such that the fish could attack but not ingest the prey. Immediately prior to the test, the fish had been provided with Tubifex ad lib for 45 min. When the fish was given the choice between pairs of tubes, one containing two Daphnia, the other containing up to 40 Daphnia, the fish showed a significant tendency to attack the two first, rather than the 40. When the choice was between two and 20, the first attacks were directed significantly more often at the 20. Other experiments showed it was the density of the swarm rather than the absolute numbers in the swarm or its volume that was the important factor. There was a critical density above which the low density (two Daphnia) was attacked more frequently than the higher density. As the distance between the pairs of tubes was increased, the number of first attacks directed at the two decreased although the
Feeding
41
total number of attacks on the two remained virtually constant. When the colour of the two and 40 Daphnia made them less conspicuous against the background, the preference for the two Daphnia decreased. This suggests that conspicuous stragglers are most at risk from stickleback predation. If the fish was made hungry by not feeding it for 1 day, then it preferred to attack the swarm of 40, although if the two were detected they were attacked at a high frequency. These observations were interpreted in terms of two factors. The denser swarms were more likely to be detected by the fish (the reaction distance was greater), but the denser swarm would also impose a greater confusion effect on the fish, making it more difficult for it to direct its attack at one particular prey (Milinski, 1977a).
0 ' 10'10 ' 10 10 10' 0 N o . Daphnia per cell Figure 4.3: Effect of Density of Daphnia on Number of First Responses (Open Bars) and Mean Number of Total Bites (Black Bars) by G. aculeatus. After Milinski (1977b)
In an extension of these experiments, the stickleback was presented with seven tubes each containing a predetermined number of Daphnia ranging from nil to 30 (Figure 4.3). This arrangement simulated a swarm of Daphnia which varied in density. The tubes were side by side,
42
Feeding
so that the fish could observe all tubes easily. The fish had not been fed for a day before the test. The first attacks were most frequently directed at the densest part of the swarm. The fish also preferred to attack the periphery of the swarm, particularly if the swarm was of equal density throughout. As the rate of attacking dropped, the fish tended to switch its attacks to the less dense regions of the swarm, although the preference for attacking the periphery was still maintained. Overall, the periphery of the swarm received more attacks than the centre, but fewer than stragglers from the swarm. Stragglers were defined as two Daphnia separated by one empty tube from the other tubes containing Daphnia. These experiments suggested that whereas a hungry fish would attack the densest part of a swarm, it would direct more and more of its attacks at stragglers from the swarm as its readiness to attack decreased (Milinski, 1977b). Experience with dense swarms did improve the performance of sticklebacks when confronted with a dense swarm (Milinski, 1979a). The 'confusion effect' was examined with a simple but ingenious experiment in which Daphnia were moved over a short distance either vertically or horizontally inside a transparent tube. Speed, direction of movement and prey density were varied independently in a controlled way, so that the stimulus complexity presented to the fish could be systematically increased (Ohguchi, 1981). A stickleback tended to show a lower responsiveness to two Daphnia if their paths crossed at right angles than to a single Daphnia, although if the two Daphnia were not moved, the fish then showed the highest responsiveness. Attacks against the cross-moving Daphnia decreased as the speed of movement increased, whereas attacks against a single Daphnia increased with the speed of movement. The frequency of attacks was greater if the two Daphnia were cross-moved out of phase with each other than if the cross-movements were in phase. If the two Daphnia differed in colour, the frequency of attacks was higher than if they were the same colour. A fish frequently attacked a piece of carrot that was cross-moved with a red Daphnia, whereas if the carrot was on its own, the frequency of attacks on the carrot was low. This suggests that the fish could discriminate between the carrot and a Daphnia less easily when the two were crossmoving. When the experimental arrangement was altered so that the Daphnia could be moved vertically and parallel to each other, the frequency of attacks was higher when the two were relatively far from each other than when they were close. If the number of Daphnia moving vertically was increased, the frequency of attacks was decreased, although if the Daphnia were not moving, an increase in number
Feeding
43
increased the attack frequency. Hungry fish attacked cross-moving Daphnia more than satiated fish, which recalls the observation that hungry fish tended to attack the denser parts of a swarm of Daphnia (Milinski, 1977b). This series of experiments suggests that a 'confusion effect' is a real phenomenon when a stickleback is attempting to attack a member of a swarm of free-swimming prey. A characteristic that makes a prey more conspicuous relative to other members of the swarm could make it more susceptible to attack, which may explain why sticklebacks tended to attack oddly coloured Daphnia more (page 34). These experiments with swarms of Daphnia show that a hungry stickleback will initially attack the densest part of the swarm, especially if it has had prior experience of such a swarm, but turns its attention to less dense parts of the swarm and to more conspicuous members of the swarm as its readiness to attack declines. The predatory stickleback selects against a tendency for the prey to stray or to differ from other members of the swarm. Apostatic selection (see below) forms an exception. Disturbances to Foraging. While foraging, the stickleback will be exposed to other stimuli, some of which may be unpleasant or potentially dangerous. A response by the fish to such stimuli will influence the pattern of the foraging sequences. The effect of an electric shock on feeding behaviour was analysed by Tugendhat (1960b,c). The fish were deprived of food for 1, 2 or 3 days and exposed to a shock classified as low, medium or high when they entered the feeding compartment to start feeding on Tubifex, or when making a feeding response. As the intensity of shock increased, the fish spent less time feeding during the hour-long test session, but more feeding responses tended to be completed and less time was spent per feeding response. These last two are characteristic of a hungry fish. At low levels of shock, the fish completed more feeding responses than unshocked fish, but at the high levels, fewer responses were completed because of the reduction in time spent feeding. Shocked fish that returned to their home compartment during a feeding session tended to spend longer there before returning to the feeding compartment than unshocked fish. If the fish were not shocked but prevented from feeding on the Tubifex by a transparent glass plate that covered the food, on subsequently being given access to the food the fish showed an increase in the time spent feeding compared with unthwarted fish. The thwarted
44
Feeding
fish tended to initiate more feeding responses by fixating a prey, but the ratio of completions was reduced. The effects of differences in the period of food deprivation prior to the test disappeared (Tugendhat, 1960c). 20r
No. Daphnia per cell Figure 4.4: Effect of Presence of Model Predator and Density of Daphnia on Number of First Responses and Mean Number of Attacks by G. aculeatus. Open Bars, Undisturbed Fish; Black Bars, Fish Disturbed by Model Kingfisher. After Milinski and Heller (1978)
The presence of a potential predator of the stickleback is likely to be a significant factor in the organisation of the stickleback's own predatory behaviour. Sticklebacks that were responding to different densities of Daphnia enclosed in seven adjacent cells were disturbed by being exposed to a moving model of the kingfisher, a known bird predator of the stickleback. Undisturbed fish initially attacked the highest density, but the disturbed fish preferentially attacked stragglers (Figure 4.4). Disturbed fish also attacked less frequently than the undisturbed fish (Milinski and Heller, 1978). In this situation, the disturbing stimulus produced behavioural effects similar to the effects of satiation, whereas the disturbance caused by a shock simulated the effect of hunger.
Optimal Foraging The behavioural analysis of foraging in the stickleback leads naturally
Feeding
45
to the question of what is the functional significance of the behavioural patterns observed. The theory of optimal foraging has been developed to predict the feeding behaviour of animals, assuming that the function of foraging behaviour is to maximise the rate of food intake. It is assumed that there is a strong correlation between feeding and the Darwinian fitness of the animal, such that natural selection tends to produce optimal foragers (Krebs, 1978). Does the organisation of the foraging behaviour of the stickleback conform to the predictions of optimal foraging?
Figure 4.5: Effect of Prey Size on the Cost of Prey Capture for S. spinachia Feeding on Mysids. Curves Estimated for Fish 70, 90 and 110mm in Length. Continuous
Line, First Prey; Broken Line, Sixth Prey. After Kislalioglu and
Gibson (1976a)
In both field and laboratory observations of S. spinachia feeding on mysids, there was a positive correlation between prey size and fish size. Experimental studies suggested that the correlation found in the field studies arose because fish of a given size had a preferred size range of prey, and different-sized prey had different swimming speeds and escape responses. As fish became satiated, they became more size-selective (Kislalioglu and Gibson, 1975). Optimal foraging theory
46
Feeding
predicts that fish should choose the most profitable prey items, so were the S. spinachia selecting prey sizes that were most profitable? Cost is inversely related to profit, and the cost of taking a prey of a given size can be measured as the time spent handling the prey (h) divided by the weight of that prey (r), i.e. hjr. For S. spinachia feeding on mysids, the handling time was determined by the ratio of prey thickness to fish mouth size. Mouth size was, in turn, related to fish length. Handling time for a prey of a given size also increased as the fish became satiated. The cost (h/r) could be predicted for a prey of given weight for a range of fish sizes and at different levels of satiation The size of prey that minimised the cost would be the optimally sized prey for a given size of fish at a given level of satiation (Figure 4.5). When the predicted optimal prey sizes were compared with the mean prey sizes taken by fish in the field, a strong positive correlation was found (Table 4.1). The theoretical analysis suggested that larger fish would have a wider range of prey sizes that were profitable, and the field data on prey sizes did indicate that the larger fish took a wider size range of prey (Kislalioglu and Gibson, 1976a). Table 4.1: Predicted O p t i m u m D i m e n s i o n s for M y s i d Prey o f S. achia
C o m p a r e d with Mean Prey Size in the Field ( F r o m
spin-
Kislalioglu
and G i b s o n , 1976a)
Fish length (mm)
70
80
90
100
110
120
2.86
3.22
3.58
3.94
Mouth size (mm)
2.14
2.50
Optimal prey size (mm)
7.5
9.0
10.5
12.0
13.5
15.0
Mean prey length in field (mm)
7.4
8.9
10.4
12.0
13.5
15.0
Optimal foraging theory was also used to interpret the behaviour of G. aculeatus feeding on Daphnia (Gibson, 1980). When sticklebacks were offered a choice between a pair of daphnids, one small (length 1.4mm), one large (length 2.4mm), the fish chose the prey that was apparently larger, irrespective of its absolute size. The smaller Daphnia would be less profitable so that a fish choosing on the basis of the apparent size of the prey would not necessarily be foraging optimally at all prey densities (O'Brien et al., 1976). The small Daphnia would appear larger to the stickleback if it were closer to the fish than the large one so that the size of its image on the retina of the fish was
Feeding
47
greater. When the fish fed on mixed populations of the two size classes, they consistently over-exploited the larger, and hence more profitable, Daphnia. If the density of the small Daphnia was increased while the density of the large Daphnia was kept constant, the proportion of small prey taken by the fish increased. But if the densities of both large and small Daphnia were increased so that the relative proportion did not change, the proportion of large prey taken did not alter. At low densities of prey, the hypothesis that the fish always took the apparently larger prey predicted successfully the proportion of large and small Daphnia taken. At high prey densities the hypothesis predicted that a higher proportion of large Daphnia would be taken, but the fish did not conform to this prediction. At high Daphnia densities, optimal foraging theory correctly predicted the observed pattern of predation on large and small Daphnia. Even at the high densities, the search times for the larger prey were never short enough for the rate of food intake to be maximised by ignoring the small Daphnia. Experiments with G. aculeatus feeding on swarms of Daphnia suggested that only hungry sticklebacks preferred to attack the densest portions of the swarm. As the fish became satiated they switched their attention from the denser parts of the swarm to the peripheral regions and to stragglers. Optimal foraging theory would suggest that the fish should always concentrate on the densest parts of the swarm, because the rate of consumption of prey would thus be maximised. An important extension to the theory of optimal foraging helps to reconcile the observations and the theory (Milinski and Heller, 1978; Heller and Milinski, 1979). Their model assumes that the stickleback is confronted with two costs. The first is the cost, measured as a reduction in fitness, associated with lack of food: in motivational terms, a 'hunger' cost. The second is the cost of an increased risk of the stickleback being predated while it is concentrating on attacking prey. This cost is higher when the fish is attacking a dense swarm because the confusion effects associated with a dense swarm demand that the fish pay more attention to the task of detecting and capturing prey and so can pay less attention to the approach of a potential predator. This can be termed the 'confusion' cost. Total costs associated with foraging are the sum of the 'hunger' and 'confusion' costs. An optimally foraging fish should seek to minimise this total cost. If a fish is very hungry, it has a high 'hunger' cost which could be minimised by feeding in the densest part of the swarm. But as the hunger is diminished, the value of each additional prey is lowered because the risk of starvation is progressively lowered and so the fish can then begin to feed in less dense
48
Feeding
parts of the swarm where the 'confusion' cost is less. The model predicts that both the density of prey that the fish selects to attack and the rate of capture of prey should decline as the fish becomes satiated. The presence of a potential predator would increase the 'confusion' cost and so should result in the threatened fish preferring to feed in less dense areas of the swarm and at a lower rate. This behaviour was indeed shown by fish exposed to a dummy kingfisher (Figure 4.4)(Milinskiand Heller, 1978). The model was also relatively successful at predicting other details of the behaviour of sticklebacks attacking swarms of Daphnia (Heller and Milinski, 1979). The model is important because it incorporates the changes in the motivational state of the animal that are brought about by the animal's behaviour into the theory of optimal foraging. When only one species of prey is present, optimal foraging can be used to predict, from the known profitabilities of the prey, which size classes should be taken and how the relative proportions of the size classes taken should change with alterations in prey density. An important extension to the theory is to incorporate the effects on foraging behaviour of the presence of more than one species of prey. When G. aculeatus were allowed to prey on Daphnia magna and larvae of the mayfly Cleon dipterum, the Daphnia were always preferred over the Qeon. The fish almost invariably caught the Daphnia at first attempt, but the success rate of capturing the 3 4 m m long Qeon was lower and decreased with an increase in the total prey density. Experiments in which the densities of Daphnia and Qeon were changed showed that the relative abundances of the two prey species were a better predictor of diet composition than the absolute density of either prey species. Although, over the densities tested, Daphnia was always the preferred prey, the degree of preference changed. First, the preference for Daphnia increased as the total prey density increased. Secondly, the preference for the relatively scarce prey increased as total prey density increased. The first result is compatible with the prediction of optimal foraging theory that as the abundance of food increases, there will be an increase in specialisation on the most profitable food items. The second result could not easily be predicted from the theory. Possibly, at high prey densities, there is a change in the profitabilities of the two prey types, perhaps as a consequence of the confusion effects created by the density of the most abundant prey. Such an effect was suggested by the reduction in the proportion of successful attacks on Qeon as the density of that species increased (Visser, 1982). The detailed analysis of foraging behaviour showed that the stickle-
Feeding
49
back had the behavioural repertoire to accept or reject prey in the context of prey density, type of prey and the motivational state of the fish. Further experiments have suggested that this behavioural repertoire is deployed in a manner that can be predicted in terms of optimal foraging theory, especially when this theory is extended to take account of the changing motivational state of the fish.
Apostatic Selection by G. aculeatus Although the experiment with oddly coloured Daphnia suggested that G. aculeatus may select prey distinguished by some oddity from its fellows, the experiments in which sticklebacks used to feeding on Tubifex were then also provided with either Drosophila larvae or enchytraeid worms indicated the importance of learning the prey characteristics. Such learning would be easier if the new prey form was encountered relatively frequently, that is, when it was the more common prey. Such a process could lead to the more common form being taken by the predator at a greater rate than suggested by its relative representation in the total prey population, because the predator has learnt its characteristics and perhaps how to capture it. In this case the predation is directed preferentially towards the common form, not the odd form, a process called apostatic selection (Clarke, 1962). Evidence for apostatic selection by sticklebacks came from experiments in which G. aculeatus were presented with populations of Asellus made up of a mixture of pale and dark individuals. When the population had an excess of pale over dark forms (40 pale, 10 dark), the pale forms were predated at a rate in excess of their relative abundance in the population, whereas when the dark form was the most abundant (40 dark, 10 pale), the dark form was over-predated. There was no differential predation when the two forms were present in equal numbers (Maskell etal., 1977). When G. aculeatus were presented with varying proportions of red and pale-yellow Daphnia, the fish showed a preference for the red form at relative proportions that varied from 80 per cent red to 20 per cent red. When the mixture contained only 12 per cent red, the yellow form was preferred. This result was unexpected because it had been anticipated that the confusion effects would be greater at high relative densities of the red form and that the fish would switch to preying upon the uncommon yellow form. Only at low densities of Daphnia did the stickleback not show a preference for the
50
Feeding
red form at high proportions of the red form (Ohguchi, 1981). Because the profitability of a prey depends on its size, and on the time required for its capture and handling, the optimal foraging strategy may be either apostatic selection or selection for oddity. Selection for oddity could be optimal when capture is relatively easy but 'confusion' costs are high. If the prey is difficult to detect or capture, yet any 'confusion' cost is not high, the optimal strategy may require apostatic selection (Visser, 1981).
Foraging in Groups Except for reproductively active males, G. aculeatus are frequently found in loose schools. If the fish are starved, they tend to spread apart and the school becomes even looser (Keenleyside, 1955). When one of the fish finds food, and particularly if it adopts the characteristic snoutdown posture to take food off the bottom, the other fish rush towards it and start searching in the same area with several fish adopting the snout-down posture. Experiments showed that a feeding fish was a stronger stimulus to a hungry fish than either food or another fish on their own (Keenleyside, 1955). This behaviour may result in the rapid aggregation of fish in a region of high food density, with an efficient exploitation of a patchy food distribution. The response of a group of G. aculeatus to a patchy distribution of food was explored by exposing six fish to two constantly renewed food patches. The food patches were formed by pipetting Daphnia into the two ends of the fish tank, but differed in the rate at which the Daphnia were added. In the first experiment, the rate of addition at one end was five times faster than at the other so the patch profitabilities were in a 5:1 ratio. The fish distributed themselves between the two patches in a ratio that closely approached a 5:1 ratio. This ratio was reached about 4 min after the start of the experiment. In a second experiment the patch profitabilities were in a ratio of 2:1 and after 9 min the profitabilities of the patches were reversed. Before the reversal, the fish again distributed themselves in a ratio that reflected the profitabilities of the patches. After the reversal, the distribution of fish approached but did not achieve the new profitability ratio. The ability of the fish to distribute themselves in ratios that reflected the density of food in the patches may represent an evolutionarily stable strategy (ESS), a foraging behaviour shown by the population which cannot be invaded by a rare mutant adopting some other feeding strategy (Milinski, 1979b, 1984).
Feeding
51
When foraging in groups, sticklebacks may compete for the same prey, so that the profitability of a prey may partly depend on an individual stickleback's ability to capture that prey in competition with another fish. Milinski (1982) showed that, for G. aculeatus, successful competitors took a higher proportion of larger Daphnia than less successful fish. When the fish Were isolated, the good competitors still preferentially attacked large Daphnia, but the poor competitors did not change their diet although no competing fish were present. This may result from the poor competitor learning which prey type it can successfully attack when in competition, but then retaining that preference. The observations of the feeding of groups of sticklebacks indicate that their repertoire of foraging behaviour allows them to forage efficiently in groups as well as when they are solitary.
Foraging in Natural Environments Laboratory experiments on foraging have used only one or two prey species and controlled prey densities, but in natural habitats the stickleback will encounter a range of potential prey species. Both the number of prey species and the overall density of prey will change seasonally and perhaps even over a day. Optimal foraging predicts that as the density of prey falls, the predator should include a wider range of prey forms in its diet. In natural habitats, a decline in prey density may also be accompanied by a reduction in the number of prey types that are available so that the stickleback is unable to increase the number of prey types -that it attacks. Although the laboratory experiments have indicated that sticklebacks have the behavioural repertoire to act as optimal foragers, only the studies on S. spinachia have suggested that they do so in the field. There is, however, considerable information on the foraging behaviour of sticklebacks in the field (Wootton, 1976). Habitat Differences in Diet Although sticklebacks inhabit a wide variety of habitats, ranging from large lakes to small streams and to coastal waters, the same prey form the bulk of the diet in all these habitats. The most prominent prey types include three microcrustacean groups, the copepods, cladocerans and ostracods; the larvae and to a lesser extent the pupae of the chironomids (midges); and the aquatic nymphs of the ephemeropterans (mayflies). Of lesser importance are oligochaetes, molluscs and algae.
52
Feeding
In brackish and coastal waters, the diet may also include larger crustaceans such as amphipods and isopods, and polychaete annelids (Lemmetyinen and Mankki, 1975; Manzer, 1976; Wootton, 1976; Hennig and Zander, 1981). These larger prey are especially important for S. spinachia (Kislalioglu and Gibson, 1975). The small size of the sticklebacks restricts the range of potential prey types and this restriction is clearly reflected in the similarity in diet over a wide range of habitats. Where two or more species share the same habitat, the overlap in diets can be considerable, which might lead to significant interspecific competition between sticklebacks for food (Chapter 8). Seasonal Changes in Diet As the abundance or availability of prey species changes during the year, the diet of the stickleback changes. An example is given by G. aculeatus living in Llyn Frongoch, a small upland reservoir in midWales. In spring and early summer, and again in late summer and early autumn, the diet was dominated by copepods and, to a much lesser extent, cladocerans. In mid-summer, the most important prey were the nymphs of ephemeropterans. Chironomid larvae were taken steadily throughout the year, whereas the pupae were taken only in late spring and summer. Algae were eaten in autumn and winter. Stickleback eggs were cannibalised in the late spring and summer (Allen and Wootton, 1984). In some populations, stickleback eggs can form a major portion of the diet in the summer (Worgan and FitzGerald, 1981b). The seasonal changes in the diet of G. aculeatus in Llyn Frongoch were similar to those of a population in a lowland stream in north-west England (Hynes, 1950) and a small pond in north-east England (Walkey, 1967). In the large Great Central Lake of Vancouver Island, the diet of G. aculeatus was dominated by copepods and cladocerans. For two species of copepods, Diaptomus and Epischura, and two species of cladocerans, Bosmina and Holopedium, their changes in abundance in the lake were reflected in changes in their importance to the stickleback (Manzer, 1976). This suggests that the seasonal changes in diet are caused by changes in the density or availability of the prey, rather than by changes in the preferences of the sticklebacks. In Llyn Frongoch, the stomach contents of G. aculeatus were heaviest in spring and late summer and lowest in winter (Allen and Wootton, 1984). The sticklebacks of Great Central Lake had their highest mean weight of stomach contents in spring and early summer. The proportion of fish with empty stomachs was lowest in spring and
Feeding
53
early summer, highest in mid-summer and again low in the autumn (Manzer, 1976). As will be discussed in Chapter 7, spring and early summer are the breeding season, a time of high energy expenditure by the sticklebacks which must be matched by a high food intake. Daily Changes in Feeding As a visual predator, the foraging behaviour of the stickleback is affected by the daily cycle of light and dark. A clear example of this was shown by G. aculeatus in Llyn Frongoch. Samples of fish were taken at 2-hourly intervals over a 24 h period at four times in a year, once in each season. Except in May, when the night was short and moonlit, there was a decline in the weight of the stomach contents during the night, indicating that feeding had either slowed down or stopped completely (Figure 4.6)
Figure 4.6: Diel Variations in Weight of Stomach Contents (mg) of G. aculeatus in Llyn Frongoch, mid-Wales. Black Bar Indicates Night-time. After Allen (1980)
54
Feeding
(Allen and Wootton, 1984). In Great Central Lake G. aculeatus showed a similar decline during the night when sampled in July. Some of the food in the stomach from fish taken late in the night was relatively fresh, indicating that some feeding during the darkness may have occurred. There was some evidence of peaks of feeding activity in the post-dawn and pre-dusk periods (Manzer, 1976). When the stomachs of G. aculeatus living in a salt-marsh pool in the St. Lawrence Estuary were sampled for a 24 h period in June, the females were found to have high rates of feeding in the post-dawn period, whereas the males did not show a peak in feeding activity (Worgan and FitzGerald, 1981b). This pattern changed when weather conditions changed (G.J. FitzGerald, pers. comm.). In Llyn Frongoch sticklebacks, there was no strong evidence for a peak of feeding associated with dawn and dusk, although in February and August, the weight of stomach contents did increase rapidly after dawn. There were few changes in the overall composition of the diet in the 24 h period; few prey species were taken only in a restricted period within the 24 h. The clearest example of such a pattern was shown in May, when nymphs of neuropteran insects were taken only at dawn and at dusk (Allen and Wootton, 1984). Experiments are needed that will define the efficiency of the foraging behaviour of sticklebacks over a range of light intensities, so that this can be related to the observed diel patterns of feeding of sticklebacks in natural populations. Effect of Fish Size on Diet The mouth size of the stickleback is related to its body size, so that larger fish have the potential of taking larger prey (Kislalioglu and Gibson, 1976a; Larson, 1976). The most rapid changes in body size take place in the first few weeks of the stickleback's life (Chapter 6), and this growth is associated with changes in the diet. As the recently hatched stickleback switches from endogenous feeding on its yolk to exogenous feeding on prey that it captures, the prey are usually the juvenile stages of cladocerans and copepods (Abdel'-Malek, 1968). In Great Central Lake, G. aculeatus less than 30mm in length ate Rotifera, and the cladocerans Bosmina and Holopedium. Larger fish tended to feed on Holopedium, the copepod Epischura, and chironomid pupae and zooplankton eggs. Fish larvae were also taken by these larger fish (Manzer, 1976). Although these effects of size can be detected, the diet of young fish rapidly becomes similar to that of older and larger fish with only relatively minor differences.
Feeding
55
Diet in Polymorphic Populations of Gasterosteus aculeatus In G. aculeatus populations containing two or more polymorphic forms (Chapters 3 and 10), differences in diet between the various morphs may be one of the factors that allow coexistence of the morphs. In the Little Campbell River in British Columbia, the low and completely plated morphs coexist during the summer in the lower reaches of the stream. The low plated morph fed primarily off the bottom of the stream, eating the aquatic larvae of insects such as coleopterans and zygopterans and a freshwater bivalve. The completely plated morph fed on swimming or surface organisms including cladocerans and a decapod crustacean. Associated with this difference in diet was a difference between the two morphs in the mean number of gill rakers which project from the inner margins of the gill arches. The completely plated morph had more gill rakers, which were longer and finer than those of the low plated morph (Hagen, 1967). Rakers probably play a role in the feeding of fish by acting as a filtering device and are usually more numerous and finer in planktivorous fishes. Paxton Lake on Texada Island, British Columbia, contains two morphs, the limnetics and the benthics (Larson, 1976). The benthics show a variable development of the lateral plates, pelvic girdle and pelvic spines (Bell, 1974), whereas the limnetics have between nine and fourteen lateral plates and a normally developed pelvic girdle and pelvic spines (Chapter 10). The gill rakers of the limnetic morph are more numerous than those of the benthic form, but the benthic form has a significantly wider mouth. The major food items of the benthics were the benthic amphipod, Gammarus lacustris, and chironomid larvae and pupae. Some benthics, mostly those less than 55mm long, also included some zooplankton in their diet. The limnetics fed mostly on cladocerans although there was some feeding on Gammarus, ostracods, chironomid larvae and mayfly nymphs. Laboratory experiments indicated that the limnetics were more successful at feeding on a cladoceran, Daphnia, than the benthics because the former spent less time having to search for and then manipulate the cladoceran. In tests where an amphipod, Hyalella, was the prey, the benthics could take larger sized prey than limnetics of the same size and spent less time manipulating the amphipod (Larson, 1976). The dietary differences between morphs noted in the Little Campbell River and Paxton Lake were associated with differences in habitat preferences, so there was a spatial as well as a dietary separation of the morphs (Chapter 10).
56
Feeding
Rates of Food Consumption Experimental Studies The most important factors that influence the quantity of food consumed by a stickleback are probably its size, physiological state and the temperature. Experimental studies have sought to determine the quantitative relationships between food consumption and these factors. Hunger. Hunger does not seem to influence the amount consumed on a daily basis. Gasterosteus aculeatus that had been deprived of food for periods that varied between 16 and 88 h did not differ in the weight of Tubifex they consumed in an 8 h period. They did differ considerably in the pattern of food intake during the 8 h period. Fish deprived for 88 h consumed 70 per cent of their total intake in the first hour, whereas fish deprived for 16 h consumed only about 35 per cent (Beukema, 1968).A similar result was obtained when sticklebacks were deprived of food for 3 days, then fed for 12 h with Tubifex, deprived for a further 12 h then fed for 12 h. The proportion of the daily food intake consumed in the first 2 h on the first day was significantly higher than the amount consumed in the equivalent period on the second day, although the total consumed in the two 12 h periods did not differ significantly (Cole, 1978). Body Weight and Temperature. The relationship between the amount of food consumed and the body weight of G. aculeatus when Tubifex was supplied ad lib was described by the relationship: l n F = a +b In W where F is weight of food, W is body weight, a and b are regression parameters and In is logarithm to base e (Wootton et al., 1980b). The value of the exponent, b, was significantly less than unity, which indicates that as the fish got heavier they consumed a relatively smaller weight of food. The same relationship between food consumption and body weight has been found in several fish species and the exponent is frequently less than unity. Ursin (1967) suggested that the maximum rate of consumption is a function of the area of the absorbing surface of the gut which is not directly proportional to body weight. The value of the parameter a in the above relationship is temperature dependent. For G. aculeatus feeding ad lib over a temperature range from 3 to 19°C, the relationship between food consumption,
Feeding
57
body weight and temperature was described by: )nF = a+bi\n.
W+
^lnT
where T is temperature and a, bj and ¿>2 a r e regression parameters. For example, when the food was Tubifex, the values of bj and b2 were 0.52 and 0.66, respectively (Wootton etal., 1980b). This relationship indicates a monotonic rise in the rate of food consumption over the temperature range used. In other species, as the upper lethal temperature is approached, the rate of food consumption then starts to decline (Elliott, 1981). This phenomenon probably also occurs in the stickleback. Although metabolic activities are in general temperature sensitive (Chapter 5), the causal mechanisms by which temperature influences the rate of food consumption are not clearly understood. Physiological State. There is some unsystematic evidence that the physiological state, particularly in relation to the phase in the cycle of sexual maturation, affects the rate of food consumption. In the period between successive spawnings of the female G. aculeatus, a period of a few days, the rate of food consumption was initially high, but in the 24 h before spawning it declined. This decline has been used as an indication that the female is about to spawn (Wootton and Evans, 1976). At the end of the spawning season, the rate of consumption by females provided with a daily ration of food equivalent to 16 per cent of their body weight showed a decline (D.A. Fletcher and R.J. Wootton, unpublished). A study of the effect of sexual maturation on the rate of food consumption and foraging behaviour would provide a valuable insight into the interactions of the two essential systems of feeding and reproduction. Control of Food Consumption Assuming that sufficient food is present in the environment, the rate of food consumption will be controlled by two systems: the demands made by the rate of metabolic activity, that is, the systemic demand, and the rate at which the food can be processed by the alimentary canal. Little is known about the mechanism by which systemic demand regulates food consumption in the stickleback. That such effects exist is indicated by the effects of deprivation on the initial rates of food consumption and also by the effects of sexual maturation. The rate at which food is processed by the alimentary canal can be measured by the rate at which the contents of the stomach are evacuated.
58
Feeding
In common with other species of fish, the rate at which the stomach of the stickleback evacuates is proportional to the weight of the contents of the stomach. This property can be described by an exponential model of stomach evacuation: dS/dt = -kS where S is the weight of stomach contents and k is the rate constant (Cole, 1978). Sibly (1981) suggests that such exponential evacuation may be an optimal tactic if there is a cost associated with processing food at the maximal rate possible for the alimentary canal. The rate of evacuation of the stomach is faster at higher temperatures (Cole, 1978). For a fish that is feeding at a rate F, the change in the weight of the stomach contents can be described by the function: dS/dt = F~kS This expression can be used to provide estimates for the rate of food consumption over a known time period if the change in the weight of stomach contents over that period and the value of k are known (Elliott and Persson, 1978). The pattern of food intake observed in sticklebacks that have been deprived of food for a pre-determined period and are then supplied with food ad lib can partially be explained by a model which assumes that the fish feed until the weight of stomach contents reaches a value of about 5-6 per cent of total body weight, and thereafter the rate of intake equals the rate of stomach evacuation (Beukema, 1968). Food Consumption in Natural Populations In laboratory experiments the rate of food consumption can be accurately measured, but the rate of food consumption in natural populations can only be estimated. Several methods have been proposed for such estimations, but all present some difficulties in application. A widely used method assumes a balanced energy budget in which the energy of the food consumed can be equated with the energy dissipated through metabolism plus the energy stored as growth or reproductive products (Chapter 5). A simple formulation of this was provided by Winberg (1956) who proposed that food consumption could be estimated from the relationship:
Feeding
59
O.SC=dB/dt+R where C is food consumed, dB/dt is growth and R is total metabolism, all measured in units of energy. Winberg argued that the total metabolism, R, could be estimated as twice the standard metabolism that could be determined in laboratory studies (Chapter 5). Subsequently, Winberg's equation has been refined, for example by Majkowski and Waiwood (1981), but the original formulation is still frequently used. Laboratory experiments yield quantitative relationships between growth and food consumption which can then be used to predict the rate of food consumption in a natural population from measurements of the latter's growth rate (Chapter 6) (Allen and Wootton, 1982a,b). Such laboratory experiments can also yield quantitative relationships between food consumption and the rate of production of faeces, which can then be used if the rate of faecal production in the natural population can be measured (Allen and Wootton, 1983). The diel pattern of change in the weight of stomach contents can also yield estimates of the rate of food consumption, especially if the rate of stomach evacuation is known (Elliott and Persson, 1978; Eggers, 1979). Food Consumption by Gasterosteus aculeatus in Llyn Frongoch These methods were used to estimate the rate of food consumption of G. aculeatus in Llyn Frongoch (mid-Wales). This is a population of slow-growing fish which reaches a weight of about 0.5-0.6g and a length of about 35mm at the end of the first year of life. The estimates for total food consumption for an average-sized stickleback in the period between the beginning of July and the end of the following June are shown in Table 4.2. Most of the estimates are reasonably close. The exception is the estimate based on the rate of faecal production. This discrepancy was partly the result of choosing an unrepresentative food for the experimental analysis of the relationship between faecal production and food consumption and partly caused by the difficulty of sampling faecal production in the field (Allen and Wootton, 1983). Estimates for the changes in the rate of food consumption within the year were made for the 1977 year class based on growth rates and the energy budget. Although the estimates for the annual consumption by the methods were similar, and there were good correlations between the methods, the energy budget tended to give higher estimates during the winter months at low water temperatures, whereas
60
Feeding
Table 4.2: Comparison of the estimates of the total annual f o o d consumption by an average-sized stickleback in Llyn Frongoch
Method of estimation
Food consumption (mg)
Growth rate
Reference
3200
Allen and Wootten (1982b)
2300 4200 2700
Allen (1980) Allen (1980) Wootton (unpublished)
3000
Allen and Wootton (1984)
3000
Allen and Wootton (1984)
12800
Allen and Wootton (1983)
Energy budget (i) Winberg I (ii) Winberg It (iii) Majkowski and Waiwood (1981) Diel samples (i) Elliott and Persson (1978) (ii) Eggers (1979) Faecal production
in late spring and summer, with higher temperatures, the growth method tended to give the higher estimates (Figure 4.7). Estimates from the diel samples using the method of Elliott and Persson (1978) showed relatively good agreement with the Winberg estimates for the same 24h periods (Table 4.3). Table 4 . 3 : Estimates of rate of f o o d consumption (mg d a y - 1 ) from four diel samplings and from Winberg's equations
Date
February
Temperature (°C)
Fish (mg)
Elliott and Persson
Winberg I
Winberg II
5.5
237
4.8
4.9
9.5
May
15.0
476
15.6
16.1
28.1
AugustSeptember
15.5
187
6.4
8.2
15.2
December
4.0
228
6.2
4.1
8.4
Feeding
61
D a y s (July 1=0) Figure 4.7: Estimates of Mean Daily Rate of Food Consumption by an Individual C. aculeatus in Llyn Frongoch between July and June. Closed Squares, Winberg II Estimate; Open Squares, Majkowski and Waiwood (1981) Method; Closed Circles, Allen and Wootton (1982b) Growth Equation
From diel samples Manzer (1976) estimated the daily ration of G. aculeatus in Great Central Lake to be 6.55 per cent of their body weight in July and 7.8 per cent in October. The daily ration of G. aculeatus in brackish water in the Baltic expressed in terms of the consumption by a fish weighing l g was estimated to be 24.2mg at 10°C, 48.3mg at 14°C, and 132mg at 18°C (Rajasilta, 1980). These estimates were based on measures of the rate of digestion. On the basis of the rate of energy expenditure measured by respiration rate, the daily ration of a stickleback weighing 4.5 g was estimated to vary from 80mg in winter to 230mg in August (Krokhin, 1957). These estimates are relatively
62
Feeding
comparable to the estimates for equivalent conditions in Llyn Frongoch. A major problem with these estimates of the rate of consumption in natural populations is their lack of precision. Nevertheless, they do indicate the order of magnitude of the rate at which the stickleback is exploiting its prey and the rate at which energy and nutrients are being made available for partitioning into survivorship, growth and reproduction. The pattern of this partitioning will depend on the environmental conditions that the fish is experiencing (Chapter 5).
Conclusion Experimental analysis of the feeding behaviour of sticklebacks had led to significant advances in the understanding of foraging tactics, both in terms of the behavioural mechanisms that are involved and in providing data, which have led to important extensions to the theory of optimal foraging. The recognition that the optimal strategy will change as the motivational state of the animal changes introduces a greater realism to the theory. The study of S. spinachia has shown that optimal foraging theory can predict some aspects of feeding in the field. As qualitative and quantitative descriptions of the diet of natural stickleback populations improve in accuracy and precision, they will provide the empirical basis for further developments of a realistic foraging theory.
5
ENVIRONMENTAL FACTORS, METABOLISM AND ENERGETICS
Introduction The rate at which food is consumed and the way in which energy from it is partitioned by the fish will depend on the environmental conditions. At one extreme, the abiotic environmental conditions may be so hostile that the stickleback is unable to maintain its metabolic and structural integrity and dies. At the other extreme, benign abiotic conditions may permit the fish to approach a maximisation of its lifetime production of offspring, if the biotic factors of food supply (Chapter 4), predation, parasitism and competition (Chapter 8) are favourable. The effects of abiotic environmental factors on the fish have been classified by Fry (1971) as lethal, controlling, limiting, masking or directive. Any factor may act in one or more of these ways. For the sticklebacks, the most important factors are probably temperature, salinity and oxygen.
Classification of Environmental Factors Lethal Factors Over the range that an abiotic environmental factor such as temperature can take, two zones can be identified. The zone of tolerance is the range over which the factor does not cause a detectable change in the survivorship of the fish. Conventionally, this is often taken to mean that the factor does not cause the death of the fish within 7 days. Outside the zone of tolerance, the metabolic processes are unable to compensate fully for the breakdown in the integrity of the fish so death follows. The zone of resistance is the range over which the factor causes the death of the fish in some measurable time period. The incipient lethal level is the value of the factor beyond which the fish can no longer live for an indefinite period. It marks the boundary between the zones of tolerance and resistance. At the outer boundary of the zone of resistance is the ultimate lethal level defining the value 63
64
Environmental Factors, Metabolism and Energetics
of the factor which causes death in a short period of time, e.g. in less than 10 min. Within the zone of resistance, an important value is the effective time, that is, the time required to bring about the lethal effect. Events within the zone of resistance are usually measured in terms of the LD50, the level of the factor that causes the death of 50 per cent of the population in a defined time period such as 48 or 96 h (Fry, 1971). The lethal level of a factor may depend on the acclimation level. This is the level of the factor the fish has experienced for a sufficiently long time to have become metabolically and physiologically adapted to that level before it encounters the potentially lethal level. Temperature. For a few species of fish, the zones of tolerance and resistance to temperature are well defined (Elliott, 1981), but this is not yet the case for the sticklebacks. For G. aculeatus and P. pungitius, their wide geographical ranges and their extensive polymorphisms suggest that details of the zones of tolerance and resistance will vary significantly between populations. Gasterosteus aculeatus that were collected in brackish water on the coast of Nova Scotia had an LDS0 of 28.8°C in tests that lasted 10000 min after they had been acclimated at 20°C and tested in a salinity of 12ppt. The lowest upper lethal temperature of 21.6°C was noted in fish that had been acclimated at 10°C and were tested in a salinity of 30ppt. The upper lethal temperature was a function both of the acclimation temperature and the salinity in which the fish were tested, but not of the acclimation salinity. Upper lethal temperature was higher at the higher acclimation temperature and lower at test salinities of Oppt and 30ppt than at 12ppt at acclimation temperatures of 10°C and 20°C. The effect of test salinity was particularly marked after acclimation at 10°C. No effect was found of fish size on the upper lethal temperature (Jordan and Garside, 1972). Specimens of G. aculeatus collected from the Columbia River on the west coast of North America had an LDS0 of 26°C in 8429 min after acclimation at 19°C. At 32°C, 50 per cent of the sample died within 2.3 min (Blahm and Snyder, 1975). Both these studies were carried out by transferring the fish directly from the acclimation temperature to the test temperature. The critical thermal maximum (CTM) for a plateless form of the stickleback collected in fresh water in southern California was determined by raising the water temperature at a rate of 1°C per 4 min from the acclimation temperature. The CTM was defined as the temperature at which opercular ventilatory movement stopped. For fish acclimated at 8°C, the CTM was 30.5°C, whereas
Environmental Factors, Metabolism and Energetics
65
after acclimation at 22.7°C, the CTM was 34.6°C. Fish size had no effect on the CTM (Feldmeth and Baskin, 1976). No comparable data are available on the lower lethal temperatures for the sticklebacks. All the species have geographical ranges which ensure that some populations regularly experience water temperatures close to 0°C. The rate of development of the eggs of sticklebacks is directly related to temperature (Heuts, 1956; Wootton, 1976). In G. aculeatus, egg survival at 25°C was poor, suggesting that this temperature is close to the lethal level for eggs. At 8°C, survival was good, but it took about 40 days for the eggs to hatch (Heuts, 1956). Salinity. Sticklebacks are euryhaline, tolerating a wide range of salinities. Lake Techirghiol, Romania, which contains a population of partially plated G. aculeatus, has a salinity that reaches 80-100ppt (Munzing, 1963). In G. aculeatus the range of tolerance varies between populations and polymorphic forms. Tolerance also varies with temperature and the physiological state of the fish. Heuts (1945) used the ability of the fish to maintain a constant osmotic pressure in their body fluids to define the range of salinity tolerances for anadromous and freshwater populations of sticklebacks in Belgium, over a temperature range from 4 to 20°C. Outside the breeding season, the freshwater population had its widest range of salinity tolerance at 10°C, a range from just above full sea water (~40ppt) to a mixture of 50 per cent distilled water and 50 per cent tap water. This range was slightly narrower at 4°C and much narrower at 20°C. (Sea water has a salinity of about 35ppt.) The anadromous form had a tolerance range between a salinity greater than sea water (45ppt) down to tap water at 4 and 10°C, but at 20°C the lower salinity tolerance was about one-third that of sea water and the upper tolerance just below the salinity of sea water (Figure 5.1). In the breeding season, the tolerance of lower salinities increased whereas the tolerance to higher salinities decreased in both populations. A comparison of the salinity tolerances of the eggs of the anadromous and freshwater populations yielded a complex picture (Heuts, 1947, 1956). At 23°C, a temperature which would be unusually high for the breeding season in Belgium, the eggs of the anadromous form tolerated higher salinities than those of the freshwater form. At 19°C, both forms had a bimodal survivorship curve, but with the salinities for maximum survivorship out of phase. The freshwater form had
66
Environmental Factors, Metabolism and Energetics
SW 1/3SW
Z
4
ife
is
¿0
Temp °C. Figure 5.1: Range of Temperature and Salinity over Which G. aculeatus Maintains a Constant Osmotic Pressure. Salinity is on a Logarithmic Scale. Vertical Shading, Anadromous Population; Horizontal Shading, Resident Freshwater Population. After Heuts (1945)
peaks of egg survival at salinities of 0.33ppt and lOppt, whereas the anadromous form had peaks of survival at 0.04ppt and 3.3ppt. At 10°C, the anadromous form had a single peak of survivorship at a salinity of about 3.3ppt, and the freshwater form had a maximum survivorship at 16ppt; the highest salinities were not tested. This series of experiments suggests, surprisingly, that the optimal salinity for egg survivorship of the freshwater form is higher than for the anadromous form over the range of temperatures likely to occur in the breeding season. A comparative review of the salinity tolerances of other stickleback species showed that C. inconstans had the lowest and S. spinachia the highest tolerance (Nelson, 1968). Culaea inconstans fed actively at salinities up to 50 per cent sea water, but at 60 per cent, most feeding had stopped, and all activity stopped at 70 per cent. Death within 24 h occurred at salinities of 70 per cent and above, although the tolerance was greater at 8°C than at 16°C. Death occurred between 5 and 15 days after C. inconstans was placed in 40 per cent sea water. Pungitius pungitius from a freshwater population stopped feeding at 80 per cent sea water at 16°C and at 90 per cent at 8°C. Apeltes quadracus, also from a freshwater population, stopped feeding at 110 and 130 per cent sea water at 16°C and 8°C, respectively.
Environmental Factors, Metabolism and Energetics
67
Even C. inconstans taken from an alkaline pond in North Dakota had a maximum salinity tolerance of 21ppt (Ahokas and Duerr, 1975), which is similar to the tolerance reported by Nelson (1968) for freshwater C. inconstans. Oxygen. In some circumstances, shallow bodies of fresh water may become deoxygenated. Warm, heavily vegetated waters, in which there are high rates of decomposition, and ice- and snow-covered lakes in winter can both reach low oxygen concentrations. The minimum oxygen concentration at which the stickleback can exist may be as low as 0.25-0.5 ppm, but normal active life at these levels is not possible (Jones, 1964). In tests in which 20 G. aculeatus were maintained in sealed 600ml beakers, the oxygen concentration when all the sticklebacks were dead was between 3.5 and 8 per cent of air saturation. In similar tests, P. pungitius, which is frequently associated with denser vegetation than the threespine stickleback, had a slightly higher toleration of low oxygen concentrations (Lewis etal., 1972). Culaea inconstans can be found in lakes that become deoxygenated in winter, such as Mystery Lake, Wisconsin, where the oxygen concentration can fall to 0-1.3 mglitre -1 . Their ability to survive these conditions depends on their use of the microlayers of oxygen associated with gas bubbles at the interface of the ice and the water. In the absence of these bubbles, confined C. inconstans suffered 50 per cent mortality in 4.3 h, whereas, in the presence of bubbles, 50 per cent mortality was reached after 21.5 h. Survivorship dropped sharply as the bubbles were depleted of oxygen. As the oxygen concentration declined, C. inconstans had a relatively low rate of respiration and increased its ventilation rate. Its activity decreased when concentrations fell to 0.5 and 0.25 mglitre"1. As the oxygen concentration of the water fell, the fish spent more time at the ice-water interface. Culaea inconstans was even observed to orientate to the flow of water from the operculum of a mudminnow that was pumping oxygenated water down through a crack (Klinger et al., 1982). Both C. inconstans and P. pungitius collected in southern Manitoba used the relatively well oxygenated surface film at the air-water interface when the oxygen content of the water declined at warm temperatures (Gee et al., 1978). Other Lethal Factors. Sticklebacks are sensitive to environmental pollutants such as heavy metals, industrial wastes and insecticides (Jones, 1964; Wootton, 1976), and are absent from heavily polluted
68
Environmental Factors, Metabolism and Energetics
waters. They will recolonise if remedial action is taken to improve water quality (Wheeler, 1979; Turnpenny and Williams, 1981). For G. aculeatus, water of pH less than 4.5 is frequently lethal so they are excluded from waters that become acidified through rainfall or poor land management. But, in the Outer Hebrides, populations are found in fresh water with a pH as low as 3.5. Such populations show a reduction in the development of lateral plates, the pelvic girdle and spines (Campbell, 1979; Giles, 1983a). Controlling Factors Controlling factors govern the metabolic rate of the fish by their influence on the state of activation of the components of the metabolic chain. A controlling factor places two bounds on the level of metabolism. It will permit a maximum rate of metabolism by its influence on the rates of chemical reactions, but it will also demand a certain minimum metabolic rate which is necessary to release the energy required for the repairs needed to keep the organism in being. The upper bound is called active metabolism and represents the maximum, sustainable rate of energy production. The lower bound is called standard metabolism and represents the minimum rate of metabolism that will maintain the steady-state integrity of the fish. It is often regarded as the metabolic rate of an unfed, resting fish. At some intermediate level is the routine rate of metabolism, which is the rate exhibited by a fish showing random, spontaneous movement. The difference between active and standard metabolism has been called the 'scope for activity' (Fry, 1971). Temperature. This is the crucial controlling factor because of its effect on the rate of chemical reactions. Fish are ectotherms living in an environment, water, which has a high specific heat and so acts as an effective heat sink. With few exceptions, fish are isothermal with their external environment (Elliott, 1981). Sticklebacks typically live in shallow waters which can show relatively rapid changes in temperature. Although the influence of temperature on the rate of chemical reactions can be described by the Arrhenius equation such that the logarithm of the rate is proportional to the reciprocal of the absolute temperature, fish do not simply respond passively to temperature changes but are able to show considerable biochemical and physiological regulation (Hochachka and Somero, 1973). Relatively little systematic analysis of the effect of temperature on the metabolic rate of sticklebacks has been published. When the rate
Environmental Factors, Metabolism and Energetics
69
respiration of G. aculeatus in a closed respirometer was measured at 7, 12.5 and 20°C (Cole, 1978), the relationship between the rate of oxygen consumption, body weight and temperature was described by: In R=a +bi In W + b2T where R is the rate of oxygen consumption, W the body weight of the fish, T the temperature, and a, bx and b2 are the parameters estimated by regression. In the experiment, the values the parameters took were: a = 2.43, by = 0.72, and b2 = 0.06, when R was in millilitres of oxygen per hour, W in grams and T in degrees Celsius. The average routine rate of metabolism for a fish weighing 0.5 g was estimated as 0.081ml 0 2 h ' 1 at 7°C, 0.113 ml 0 2 h" 1 at 12.5°C and 0.177 ml 0 2 h" 1 at 20°C. Studies on other fish species suggest that both active and standard rates of metabolism also increase with temperature, although at high temperatures a shortage of oxygen may start to limit metabolism (Brett, 1979). The rate of food consumption also increases with a rise in temperature, at least up to some optimal temperature at which the rate is maximal (Chapter 4). Limiting Factors A limiting factor operates by restricting the supply of substrates or the removal of products in the metabolic chain so that the rate of metabolism is less than that permitted by the levels of the controlling factors. The crucial limiting factor for fish is most likely to be the biotic factor of food supply, but under some circumstances the supply of oxygen may be inadequate to support the rate of metabolism permitted by the environmental temperature. Oxygen. Low oxygen concentration has been shown to restrict both activity and growth in some fish (Brett, 1979). Unplated southern Califomian G. aculeatus showed a rapid decline in their respiration rate at 18.6°C when the oxygen concentration of the water fell below 2ppm. Above this concentration, the rate of respiration was independent of oxygen consumption, but at concentrations below 7.4ppm, the ventilation rate increased to maintain the respiration rate at a constant level (Feldmeth and Baskin, 1976). The energy required to maintain the high ventilation rates at the low oxygen concentrations may be diverted from growth or other activities (Feldmeth and Baskin, 1976). Culaea inconstans showed a decrease in locomotory activity when the oxygen concentration fell below about 1 mglitre -1 in water
70
Environmental Factors, Metabolism and Energetics
at 2.5-4.0°C (Klinger et al., 1982). At 16°C, C. inconstans started to show a decline in the ventilation rate and in activity when the partial pressure of oxygen in the water dropped to 34.1 Torr (saturation partial pressure was approximately 150Torr). At 18.7Torr, 50 per cent of the fish were using the surface film. A decline in activity was shown by P. pungitius at 58.7Torr, and at 49.9Torr, 50 per cent of the fish were using the surface film (Gee et al., 1978). Masking Factors Masking factors exert an additional metabolic demand on the fish at particular levels of the controlling and limiting factors. This metabolic demand reduces the energy that can be partitioned into other components such as growth or behavioural activity. The metabolic demand represents the energy cost the fish has to pay in order to maintain either a constant internal or external environment in the face of the masking factor (Fry, 1971). Salinity. Gasterosteus aculeatus has the ability to maintain a relatively constant osmotic pressure and ionic concentration in its body fluids over a wide range of external salinities (Heuts, 1945). The rate of oxygen consumption of the completely plated and partially plated morphs was significantly higher in sea water (35ppt salinity) than in fresh water at both 4 and 20°C, whereas the low plated form did not show a significant difference in the rate of respiration. This suggests that the osmoregulation of the completely and partially plated morphs in sea water requires the expenditure of energy whereas that of the low plated form uses a mechanism that does not require an expenditure of energy detectable by respirometry. The low plated morph osmoregulated by altering the level of free amino acids in the body fluids (Gutz, 1970). Its respiration rate showed no change over a range of salinities from 1.7 to 27.2ppt at 12°C after acclimation for 5 weeks (S. McGibbon and R.J. Wootton, unpublished), which also suggests that salinity does not act as a masking factor for this morph. During the breeding season, the active osmoregulation of the completely and partially plated morphs became less effective and they adopted the mechanism of the low plated form. The mechanism of the low plated form was ineffective in sea water at low water temperatures (Gutz, 1970). These differences in osmoregulation are reflected in the global distributions of the three morphs (Chapter 10). Information on the effect of salinity on the growth rates of the polymorphic forms of the stickleback would be of considerable interest.
Environmental Factors, Metabolism and Energetics
71
Current. If a fish wishes to maintain its position in a current, it must expend energy on mechanical work. No information is available on the effect of current as a masking factor in the stickleback, but the fish tends to choose areas of still or relatively slow-moving water and so may minimise the energy costs of maintaining spatial position. Directive Factors Such factors allow or require a response by the fish which is directed in relation to a gradient in the factor in space or time. These directed responses are important because they allow the fish some choice over the environment it experiences in addition to its biochemical and physiological capacities for adaptation to environmental circumstances. Sticklebacks respond to gradients in temperature and salinity. Temperature. When G. aculeatus that had been collected from intertidal pools on the coast of Nova Scotia were tested in a vertical temperature gradient that ranged from 5°C at the bottom to 28°C at the top, their preferred temperature depended on the acclimation temperature and the salinity of the water (Garside et al., 1977). At an acclimation temperature of 5°C, the modal preferred temperature in sea water was 13°C, but in fresh water 10°C. With acclimation at 15°C, the equivalent modes were at 17°C and 16°C, and with acclimation at 25°C, the modes were 20°C in sea water and 16°C in fresh water. The final preferendum, defined as the value at which the preferred temperature is equal to the acclimation temperature, was estimated at 16°C in fresh water and 18°C in sea water. Gasterosteus aculeatus collected in Oslofjorden, Norway, showed much lower temperature preferences when tested in a vertical gradient that consisted of three linked compartments which differed in temperature by 5°C. Irrespective of the acclimation temperature, which varied from 5 to 20°C, the fish preferred temperatures lower than 11°C and in most cases between 4 and 8°C (R0ed, 1979). This preferred range was less than the temperatures in which the fish were naturally living for most of the year, suggesting that the result may be an experimental artefact. Salinity. In a comprehensive series of experiments on the salinity preferences of an anadromous population of G. aculeatus which consisted of a mixture of the completely, partially and low plated morphs, Baggerman (1957) used a 100 cm-long trough that was divided into four compartments. Two adjacent compartments contained fresh water, the next contained water with a salinity of 7.8-8.6ppt, and the
72
Environmental Factors, Metabolism and Energetics
fourth compartment had a salinity of 11.6-13.4 ppt. A thin film of fresh water connected the compartments and allowed the fish to travel between them. In this apparatus, the fish showed salinity preferences which changed during the year. In the early months of the year, there was a significant preference for fresh water, but from June onwards the preference was for salt water. These changes in preference correspond to the reproductive cycle. In late winter, the fish move from the sea to breed in fresh water, with the surviving adults and young-of-the-year moving back to the sea after the spring breeding season. The young fish showed a preference for salt water within 2 months of hatching. Experiments showed that the changes in preference were correlated with the onset of sexual maturity and the end of the breeding season, but were not dependent on the presence of developing gonads. A sharp temperature rise in November did not produce a change in the preference for salt water, whereas the same rise in the early months of the year produced a change to a preference for fresh water within a few days. Fish treated with the thyroid hormone, thyroxine, showed a change from a preference for salt water to one for fresh water within 3 to 5 days. Thiourea, an inhibitor of the thyroid, caused a preference for fresh water to change to one for salt water, again within a few days. This study indicates the close relationship between the directive response to the salinity gradient and the physiological state of the stickleback. The fish does not simply respond in a passive and non-adaptive way to the environmental conditions. Other Factors. When presented with a choice between water at pH 6.8 and acidified water, G. aculeatus avoided water with a pH of less than 5.8. They also avoided water with a pH greater than about 11. Over the pH range 6.0-11.0, the fish appeared to be indifferent. When given a choice between water with a high or a low oxygen concentration, the fish crossed the boundary into the low-oxygen water without hesitation, but then became agitated and so eventually returned to the well oxygenated zone where they tended to become still. At 13°C, the escape from the poorly oxygenated zone was relatively rapid, with oxygen concentrations of less than 3ppm. At low temperature the response was similar but took longer to develop. In this situation, the fish did not seem to recognise the presence of a gradient and only escaped from a potentially harmful situation by random movements (Jones, 1964).
Environmental Factors, Metabolism and Energetics
73
The Energy Budget Introduction Although the pathways by which the energy intake is partitioned between outputs Will depend on the environmental conditions and the physiological state of the fish, the biochemical, physiological and behavioural mechanisms that allow the stickleback to select and adapt to the environment probably maintain the fish within its zones of tolerance except in exceptional circumstances. Other causes of death, such as predation, parasitism and the costs associated with reproduction, probably far outweigh in importance the deaths caused by abiotic environmental factors.
Figure 5.2: Model of Energy Partitioning by an Individual Fish. After Warren and Davis (1967)
A convenient framework for the discussion of the energy partitioning of a stickleback is the bioenergetics model of Warren and Davis (1967) shown in Figure 5.2. The energy budget for a defined time period is defined as: C=F+U+P+R where C is the energy content of the food consumed, F is the energy content of faeces, U is the energy content of excretory products, P
74
Environmental Factors, Metabolism and Energetics
is the change in the total energy content of the body including any reproductive products released during the period, and R is the total energy of metabolism. R can be subdivided into three components, Rs, Ra and Rtj, where Rs is the standard metabolism or the rate of energy expenditure of a resting unfed fish, is the additional rate of energy expenditure of an active fish, and is the additional rate of energy expenditure associated with the ingestion and digestion of food and is usually called the apparent specific dynamic action (SDA). A weakness of the Warren and Davis model of energy partitioning is that it does not suggest that the fish can exert control over the details of the partitioning. Models which do incorporate such control have been described by Hubbell (1971) and Calow (1973) but such models have not yet been developed for fish. Survival, growth and reproduction do not depend only on the rate at which energy is gained and expended. They also depend on the food providing vitamins and minerals, which do not yield energy but are essential for the operation of the metabolic processes. The study of energy budgets ignores these components of the diet and so can be misleading. Unfortunately, little is known about the requirements of sticklebacks for such components (Love, 1980). Components of the Energy Budget Faecal Losses (F). Not all the food consumed is digested, so the fish loses some energy in the form of faeces, which will also include mucus, cells sloughed off the wall of the alimentary canal and other secretions (Elliott, 1979). Absorption efficiency is measured as the proportion of the energy in the food that is not lost as faeces. In terms of the Warren and Davis model, the efficiency is defined as: (C — F)/C, expressed as a percentage. Ideally, C and F must be measured in energy units, but an indication of the absorption efficiency can be obtained if they are measured in terms of dry weight. Fed to satiation twice a day with Tubifex, G. aculeatus had absorption efficiencies of between 86 and 96 per cent, where C and F were measured in energy units. The faeces had a mean energy content of 16.7kJ g - 1 dry weight (Cole, 1978). The absorption efficiency of fish supplied with unlimited Tubifex was estimated as 85 per cent and the energy content of the faeces as 18.3kJ g"1 dry weight (Walkey and Meakins, 1970). In both studies, the small quantities of faeces produced by the fish meant that the energy content was measured on only a few samples. When F and C were measured in dry weights with the fish fed enchytraeid worms,
Environmental Factors, Metabolism and Energetics
75
the absorption efficiency was 86-96 per cent. In this experiment, absorption efficiency was higher at higher temperatures and was also higher for heavier fish, but was not related to the size of the ration consumed by the fish (Allen and Wootton, 1983). These experiments probably overestimate the absorption efficiency because any soluble components of the faeces were not measured. The natural diet of sticklebacks usually includes a high proportion of prey with chitinous exoskeletons, which are not readily digested, so the absorption efficiencies are likely to be lower. Gasterosteus aculeatus collected in the field produced a much higher weight of faeces in the 24 h after capture than fish produced in 24 h when fed to satiation with enchytraeid worms (Allen and Wootton, 1983). Excretory Products (U). The stickleback also loses energy in the form of its excretory products. In fishes, ammonia is the major excretory product and each milligram of ammonia-nitrogen lost represents 24.85 J (Elliott, 1979). There has been no detailed study of the rate of production of excretory products by the stickleback, but it has been suggested that energy losses from this source can represent 3-5 per cent of the energy intake (Walkey and Meakins, 1970). Studies on the trout, Salmo trutta, suggest that at high temperatures and low rations such losses can reach up to 15 per cent of the intake (Elliott, 1979). Assimilation efficiency measures the proportion of the energy of the food consumed that remains after the faecal and excretory losses have been taken into account. It is defined as [C- (F+U)]/C. In studies on energy budgets of fish it is frequently taken to be 80 per cent, but it is probably a function of temperature, ration, composition of the food and other factors (Elliott, 1979). The energy left after faecal and excretory losses have been taken into account can be regarded as physiologically useful energy. It is used for metabolism and stored in the form of new tissue as somatic growth and reproductive products. Energy of Metabolism (Rj. The rate at which energy is being used in metabolism can be measured as the rate of heat production by the fish, or, more conveniently, by the rate of respiration. Each milligram of oxygen consumed is equivalent to an energy expenditure of about 13.56J (Elliott and Davison, 1975). In this way the rates of oxygen consumption by the fish under various environmental conditions can be converted to rates of energy expenditure. The rate of respiration is a function of the body weight of the fish of the form:
76
Environmental Factors, Metabolism and Energetics R
or in the linear form: In/? = lna +b In W where R is the rate of respiration, W is the body weight, and a and b are parameters (cf. page 56). In fishes, the exponent, b, is typically less than unity and often approximately 0.8. For the measurement of the rate of respiration, the fish has to be enclosed in a respirometer, which typically provides only a restricted volume, so that although relatively accurate measurements can be obtained, their relevance for a free-ranging fish are more difficult to judge. The enclosed nature of the respirometer may stress the fish, so that the rates of energy expenditure are unusually high unless the fish is given time to become used to the conditions. In a study of the effects of activity on the metabolic rate of G. aculeatus, single fish were enclosed in a respirometer containing 500ml of aerated water and then the respirometer was sealed and the decline in the oxygen content of the water was measured. In the first experiment, the fish were allowed to show spontaneous activity swimming at their own speed: this provided a measurement of the routine rate of metabolism. In a second experiment, the fish were made to swim against a current created by a magnetic stirrer. The speed of the current was increased until the fish could just maintain a constant position: this yielded a measurement of the maximum rate of metabolism. Eventually, the fish tired and were passively swept round by the current: this provided a measurement of the minimal rate of metabolism. The experiments were carried out at 15°C, but on fish collected in either February or August. In all conditions, the weight exponent, b, in the relationship between rate of respiration and body weight was less than unity, indicating that the heavier fish had a relatively slower rate of energy expenditure when this was expressed per gram body weight. The maximum and minimum rates were significantly higher in August than in February, but the routine rates did not differ significantly. In February, the maximum rate was 4.5 times the minimum, and in August, the ratio of maximum to minimum was 3.8 (Meakins, 1975; Meakins and Walkey, 1975). The minimum rate could be regarded as an estimate of Rs and the difference between the maximum and minimum rates as an estimate of R a for the maximum level of activity, although the artificial environment used in the experiments
Environmental Factors, Metabolism and Energetics
77
makes it difficult to extrapolate these results to other situations. A significant correlation between swimming speed and the rate of oxygen consumption was shown by G. aculeatus kept at 13-14°C (Lester, 1971). At a swimming speed of one fish length per second, the rate of oxygen consumption was about l.Omlh" 1 g"1 dry weight, but this rate was doubled at twice the swimming speed. The rate of oxygen consumption increases after the consumption of a meal up to a peak and then declines back to the pre-feeding rate, indicating an increase in the rate of energy expenditure associated with food processing. In fish this energy expenditure, the apparent specific dynamic action, typically accounts for about 9-26 per cent of the energy of the food being processed (Jobling, 1981). The causes of this increase in energy expenditure are not clearly understood. A small part represents the energy required to move the food through the alimentary canal and the processes associated with uptake of food across the wall of the gut, but the major component is probably associated with the metabolism of proteins and amino acids (Jobling, 1983b). An increase in the rate of oxygen consumption after feeding has been found in G. aculeatus, but the values of over a range of rations and temperatures are not known (E.M. Lewis, unpublished). Production (Pj. Physiologically useful energy that is not expended in metabolism is stored in the form of new tissue. Some of this new tissue increases the body mass of the fish, and some consists of reproductive products, typically eggs or sperm. In the case of the male stickleback, it also includes the glue secreted by the kidneys during nest construction. Thus, production can be divided into somatic production, Ps, and reproductive production, Pr These two components are discussed in Chapters 6 and 7. Energy Budgets for the Sticklebacks It is impossible to measure simultaneously all the components of the energy budget of a fish, so the budget as a whole is not of interest. Living animals are not good subjects for a critical test of the laws of thermodynamics; rather, the laws are invoked with the assumption that the energy income will be exactly balanced by the expenditure. The important problem is to determine how the components change as environmental factors and the physiological state of the fish change. The best example of this approach is the analysis of the energy budget of the brown trout in relation to temperature, body weight and food supply (Elliott, 1979). In the stickleback, more attention has been paid
78
Environmental Factors, Metabolism and Energetics
to the problem of the partitioning of energy between somatic and reproductive production (Chapters 6 and 7), but some incomplete budgets have been estimated. The effects of body weight and temperature on some components of the budget for G. aculeatus are shown in Table 5.1. The fish were not in reproductive condition and were fed to satiation with Tubifex twice a day. Regression analysis was used to obtained predictions of each of the budgetary components for selected fish weights. Several components were not estimated, including the excretory products (U),Rd, and the energy costs of activity above the routine level. Table 5.1: Effect of Temperature and Body Weight on Some Components of the Energy Budget of G. aculeatus (Cole, 1978) Temperature ( C)
Fish weight (g)
Energy content offish (kJ)
Food consumed (kJ day - 1 )
Food absorbed (kJ day" 1 )
Routine respiration (kJ day" 1 )
Growth (kJ day )
7
0.500 1.000 1.500
2.146 4.494 6.782
0.132 0.212 0.262
0.126 0.200 0.245
0.039 0.053 0.064
0.007 0.020 0.025
12.5
0.500 1.000 1.500
2.146 4.494 6.782
0.191 0.307 0.401
0.180 0.285 0.371
0.057 0.097 0.131
0.024 0.034 0.045
20
0.500 1.000 1.500
2.146 4.494 6.782
0.258 0.414 0.542
0.240 0.384 0.496
0.077 0.157 0.237
0.028 0.048 0.062
Partial energy budgets for G. aculeatus infected with plerocercoids of the cestode Schistocephalus solidus were compared with those for uninfected fish (Walkey and Meakins, 1970; Meakins, 1974a). The fish were kept at 15°C and fed an ad lib ration of Tubifex. Absorbed energy was estimated as 80 per cent of the energy of the consumed food, and metabolic expenditure was estimated from the rate of oxygen consumption shown by a fish swimming at the maximum rate. The average values for the components of the budget were for the unparasitised fish: assimilated energy, 0.563 kJday - 1 ; respiration, 0.242 kJday - 1 ; growth, 0.052 kJday - 1 . In parasitised fish the average values were: assimilated energy, 0.517 kJday" 1 ; respiration, 0.293 kJday - 1 ; growth, 0.046 kJday" 1 . For the parasitised fish, the growth component consisted of the growth of the fish and the parasites (Chapter 8).
Environmental Factors, Metabolism and Energetics
!£l O Ii) (N oo oo oo oo f ?
roo o o o o
NC CO o co f- C M in 5 8 o o oo oo oo o o P o o o o
oM C o o
e 3o 5
> C M M MC O 9 O oCO o CO C £< o 0 " O o o o o o o o o o o o o o " -R, J d o o o' o o o o o o o" o" o o o o"
(A
9I a >
CN C M CN CO CO o o o p o oo oo oo oo o o o o o" o o d o o o"
CN CN CN CN C O co oo oo 8 8 ooM C 8 oo oo o o o o o 0
OOOOdT-OOlDr-O) co cn m rin ^. co to I» 00 ^ n t f n n i t p h S 0 0 0 0 0 0 0 — o o o o o o o o o d o d o o" o o" o o o o o o o o
E
o in in p in o o p co o> ID to co iri r-°
1 E c/5 O Z O
ao3 a.
o o in in o in o Ö (D (I) ID ID 4 o
•C
•a .o >. C | £ 5 E -5 = 5 va> £ -O != s O a, ra D.